Embed Size (px)
Citation preview
Transmission delay (per packet) = LR = amount of time to transmit each packet onto linkPropagation delay (per bit) = ds = amount of time for a single bit to transit the linkAnd for this simplified case assume no queuing delay or processing delay in routers
End-to-End Delay
First note that there is no queue at each router so packets are transmitted onto links 2 3 and 4 with no delay That is as soon as last bit of each packet arrives in a router it begins transmission onto next link
For example the first packet begins transmission onto Link 2 after its end-to-end delay on Link 1 (LR + ds) and finishes transmission onto Link 2 at time (LR + ds) + LR The second packet arrives at the first router at time LR (wait time on first packet at source host) + LR + ds (its end-to-end delay on Link 1) or LR + LR + ds THE SAME TIME THAT THE PRIOR PACKET HAS COMPLETED TRANSMISSION onto the next link
Link 1 Link 4Link 3Link 2
Link 1 Link 4Link 3Link 2
You should convince yourself that in this simplified case since L is the same for each packet and link characteristics R d and s are the same for each link that this behavior holds for each packet at each router regardless of number of packets or links
Now end-to-end delay for all 4 packets through this network is simply the time that the last bit of last packet packet 4 arrives at Host 2
End-to-End Delay
Note that packet 4 cannot begin transmission until the 3 packets ldquoin front of itrdquo have completed transmission or LR (packet 1) + LR (packet 2) + LR (packet 3) = 3 LR
Link 1 Link 4Link 3Link 2
Link 1 Link 4Link 3Link 2
Note also that by prior argument this packet proceeds through the network unimpeded by any queuing delay at the intervening routers so its total end-to-end delay is LR + ds for each link
Link 1 Link 4Link 3Link 2
Time 3LR
Time 3LR + LR+ds + LR+ds + LR+ds + LR+ds
So Total end-to-end delay = 3LR + 4LR + 4 ds Generalized for n packets and k links
(n-1) LR + k (LR + ds)
Transport Layer 3-3
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-4
TCP Overview RFCs 79311221323 2018 2581
connection-oriented ie requires setup in
end-systems before data can be exchanged
ldquohandshakingrdquo (exchange of control messages) initializes sender amp receiver states (per-connection variables) before data exchange
flow controlled sender will not
overwhelm receiver by sending data too fast
point-to-point one sender one receiver
reliable in-order byte steam no ldquomessage boundaries
rdquo pipelined
TCP congestion and flow control set window size
full duplex data bi-directional data flow
in same connection MSS maximum segment
size
TCP Logical End-to-End Connection
Transport Layer 3-5
socketdoor
TCP send buffer
TCP receive buffer
socketdoor
segment
application processwrites data
application processwrites data
segment
a TCP connection is point-to-point only between a single sender and a single receiver
Multicast with TCP is not possible
a TCP connection is point-to-point only between a single sender and a single receiver
Multicast with TCP is not possible
Transport Layer 3-6
TCP segment structure
source port dest port
32 bits
applicationdata
(variable length)
sequence number
acknowledgement number
receive window
URG data pointerchecksum
FSRPAUheadlen
notused
options (variable length)
URG urgent data (generally not used)
ACK ACK valid
PSH push data now(generally not used)
RST SYN FINConnection mgmt
(setup teardowncommands)
bytes rcvr willingto accept
countingby bytes of data(not segments)
Internetchecksum
(as in UDP)
32-bit wordsin header
Transport Layer 3-7
TCP seq numbers ACKssequence numbers
byte stream ldquonumberrdquo of first byte in segmentrsquos data
acknowledgementssequence of next byte expected from other side
cumulative ACKQ how receiver handles out-of-order segmentsA TCP spec doesnrsquot say - up to implementer
A SACK option possible per RFC 2018
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
incoming segment to sender
A
sent ACKed
sent not-yet ACKed(ldquoin-flightrdquo)
usablebut not yet sent
not usable
window size N
sender sequence number space
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
outgoing segment from sender
Transport Layer 3-8
TCP seq numbers ACKs
Usertypes
lsquoCrsquo
host ACKsreceipt
of echoedlsquoCrsquo
host ACKsreceipt oflsquoCrsquo echoesback lsquoCrsquo
simple telnet scenario
Host BHost A
Seq=42 ACK=79 data = lsquoCrsquo
Seq=79 ACK=43 data = lsquoCrsquo
Seq=43 ACK=80
Transport Layer 3-9
TCP round trip time timeoutQ how to set TCP
timeout value longer than RTT
but RTT varies too short premature
timeout unnecessary retransmissions
too long slow reaction to segment loss
Q how to estimate RTT
SampleRTT measured time from segment transmission until ACK receipt ldquobest practicerdquo uses TCP
timer option per RFC 1323 ignore retransmissions
SampleRTT will vary so we want estimated RTT to be ldquosmootherrdquo average several recent
measurements not just current SampleRTT
Transport Layer 3-10
RTT gaiacsumassedu to fantasiaeurecomfr
100
150
200
250
300
350
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
RTT
(mill
isec
onds
)
SampleRTT Estimated RTT
EstimatedRTT = (1- )EstimatedRTT + SampleRTT exponential weighted moving average influence of past sample decreases
exponentially fast typical value = 0125
TCP round trip time timeout
RTT
(mill
iseco
nds)
RTT gaiacsumassedu to fantasiaeurecomfr
sampleRTT
EstimatedRTT
time (seconds)
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Now end-to-end delay for all 4 packets through this network is simply the time that the last bit of last packet packet 4 arrives at Host 2
End-to-End Delay
Note that packet 4 cannot begin transmission until the 3 packets ldquoin front of itrdquo have completed transmission or LR (packet 1) + LR (packet 2) + LR (packet 3) = 3 LR
Link 1 Link 4Link 3Link 2
Link 1 Link 4Link 3Link 2
Note also that by prior argument this packet proceeds through the network unimpeded by any queuing delay at the intervening routers so its total end-to-end delay is LR + ds for each link
Link 1 Link 4Link 3Link 2
Time 3LR
Time 3LR + LR+ds + LR+ds + LR+ds + LR+ds
So Total end-to-end delay = 3LR + 4LR + 4 ds Generalized for n packets and k links
(n-1) LR + k (LR + ds)
Transport Layer 3-3
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-4
TCP Overview RFCs 79311221323 2018 2581
connection-oriented ie requires setup in
end-systems before data can be exchanged
ldquohandshakingrdquo (exchange of control messages) initializes sender amp receiver states (per-connection variables) before data exchange
flow controlled sender will not
overwhelm receiver by sending data too fast
point-to-point one sender one receiver
reliable in-order byte steam no ldquomessage boundaries
rdquo pipelined
TCP congestion and flow control set window size
full duplex data bi-directional data flow
in same connection MSS maximum segment
size
TCP Logical End-to-End Connection
Transport Layer 3-5
socketdoor
TCP send buffer
TCP receive buffer
socketdoor
segment
application processwrites data
application processwrites data
segment
a TCP connection is point-to-point only between a single sender and a single receiver
Multicast with TCP is not possible
a TCP connection is point-to-point only between a single sender and a single receiver
Multicast with TCP is not possible
Transport Layer 3-6
TCP segment structure
source port dest port
32 bits
applicationdata
(variable length)
sequence number
acknowledgement number
receive window
URG data pointerchecksum
FSRPAUheadlen
notused
options (variable length)
URG urgent data (generally not used)
ACK ACK valid
PSH push data now(generally not used)
RST SYN FINConnection mgmt
(setup teardowncommands)
bytes rcvr willingto accept
countingby bytes of data(not segments)
Internetchecksum
(as in UDP)
32-bit wordsin header
Transport Layer 3-7
TCP seq numbers ACKssequence numbers
byte stream ldquonumberrdquo of first byte in segmentrsquos data
acknowledgementssequence of next byte expected from other side
cumulative ACKQ how receiver handles out-of-order segmentsA TCP spec doesnrsquot say - up to implementer
A SACK option possible per RFC 2018
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
incoming segment to sender
A
sent ACKed
sent not-yet ACKed(ldquoin-flightrdquo)
usablebut not yet sent
not usable
window size N
sender sequence number space
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
outgoing segment from sender
Transport Layer 3-8
TCP seq numbers ACKs
Usertypes
lsquoCrsquo
host ACKsreceipt
of echoedlsquoCrsquo
host ACKsreceipt oflsquoCrsquo echoesback lsquoCrsquo
simple telnet scenario
Host BHost A
Seq=42 ACK=79 data = lsquoCrsquo
Seq=79 ACK=43 data = lsquoCrsquo
Seq=43 ACK=80
Transport Layer 3-9
TCP round trip time timeoutQ how to set TCP
timeout value longer than RTT
but RTT varies too short premature
timeout unnecessary retransmissions
too long slow reaction to segment loss
Q how to estimate RTT
SampleRTT measured time from segment transmission until ACK receipt ldquobest practicerdquo uses TCP
timer option per RFC 1323 ignore retransmissions
SampleRTT will vary so we want estimated RTT to be ldquosmootherrdquo average several recent
measurements not just current SampleRTT
Transport Layer 3-10
RTT gaiacsumassedu to fantasiaeurecomfr
100
150
200
250
300
350
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
RTT
(mill
isec
onds
)
SampleRTT Estimated RTT
EstimatedRTT = (1- )EstimatedRTT + SampleRTT exponential weighted moving average influence of past sample decreases
exponentially fast typical value = 0125
TCP round trip time timeout
RTT
(mill
iseco
nds)
RTT gaiacsumassedu to fantasiaeurecomfr
sampleRTT
EstimatedRTT
time (seconds)
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-3
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-4
TCP Overview RFCs 79311221323 2018 2581
connection-oriented ie requires setup in
end-systems before data can be exchanged
ldquohandshakingrdquo (exchange of control messages) initializes sender amp receiver states (per-connection variables) before data exchange
flow controlled sender will not
overwhelm receiver by sending data too fast
point-to-point one sender one receiver
reliable in-order byte steam no ldquomessage boundaries
rdquo pipelined
TCP congestion and flow control set window size
full duplex data bi-directional data flow
in same connection MSS maximum segment
size
TCP Logical End-to-End Connection
Transport Layer 3-5
socketdoor
TCP send buffer
TCP receive buffer
socketdoor
segment
application processwrites data
application processwrites data
segment
a TCP connection is point-to-point only between a single sender and a single receiver
Multicast with TCP is not possible
a TCP connection is point-to-point only between a single sender and a single receiver
Multicast with TCP is not possible
Transport Layer 3-6
TCP segment structure
source port dest port
32 bits
applicationdata
(variable length)
sequence number
acknowledgement number
receive window
URG data pointerchecksum
FSRPAUheadlen
notused
options (variable length)
URG urgent data (generally not used)
ACK ACK valid
PSH push data now(generally not used)
RST SYN FINConnection mgmt
(setup teardowncommands)
bytes rcvr willingto accept
countingby bytes of data(not segments)
Internetchecksum
(as in UDP)
32-bit wordsin header
Transport Layer 3-7
TCP seq numbers ACKssequence numbers
byte stream ldquonumberrdquo of first byte in segmentrsquos data
acknowledgementssequence of next byte expected from other side
cumulative ACKQ how receiver handles out-of-order segmentsA TCP spec doesnrsquot say - up to implementer
A SACK option possible per RFC 2018
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
incoming segment to sender
A
sent ACKed
sent not-yet ACKed(ldquoin-flightrdquo)
usablebut not yet sent
not usable
window size N
sender sequence number space
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
outgoing segment from sender
Transport Layer 3-8
TCP seq numbers ACKs
Usertypes
lsquoCrsquo
host ACKsreceipt
of echoedlsquoCrsquo
host ACKsreceipt oflsquoCrsquo echoesback lsquoCrsquo
simple telnet scenario
Host BHost A
Seq=42 ACK=79 data = lsquoCrsquo
Seq=79 ACK=43 data = lsquoCrsquo
Seq=43 ACK=80
Transport Layer 3-9
TCP round trip time timeoutQ how to set TCP
timeout value longer than RTT
but RTT varies too short premature
timeout unnecessary retransmissions
too long slow reaction to segment loss
Q how to estimate RTT
SampleRTT measured time from segment transmission until ACK receipt ldquobest practicerdquo uses TCP
timer option per RFC 1323 ignore retransmissions
SampleRTT will vary so we want estimated RTT to be ldquosmootherrdquo average several recent
measurements not just current SampleRTT
Transport Layer 3-10
RTT gaiacsumassedu to fantasiaeurecomfr
100
150
200
250
300
350
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
RTT
(mill
isec
onds
)
SampleRTT Estimated RTT
EstimatedRTT = (1- )EstimatedRTT + SampleRTT exponential weighted moving average influence of past sample decreases
exponentially fast typical value = 0125
TCP round trip time timeout
RTT
(mill
iseco
nds)
RTT gaiacsumassedu to fantasiaeurecomfr
sampleRTT
EstimatedRTT
time (seconds)
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-4
TCP Overview RFCs 79311221323 2018 2581
connection-oriented ie requires setup in
end-systems before data can be exchanged
ldquohandshakingrdquo (exchange of control messages) initializes sender amp receiver states (per-connection variables) before data exchange
flow controlled sender will not
overwhelm receiver by sending data too fast
point-to-point one sender one receiver
reliable in-order byte steam no ldquomessage boundaries
rdquo pipelined
TCP congestion and flow control set window size
full duplex data bi-directional data flow
in same connection MSS maximum segment
size
TCP Logical End-to-End Connection
Transport Layer 3-5
socketdoor
TCP send buffer
TCP receive buffer
socketdoor
segment
application processwrites data
application processwrites data
segment
a TCP connection is point-to-point only between a single sender and a single receiver
Multicast with TCP is not possible
a TCP connection is point-to-point only between a single sender and a single receiver
Multicast with TCP is not possible
Transport Layer 3-6
TCP segment structure
source port dest port
32 bits
applicationdata
(variable length)
sequence number
acknowledgement number
receive window
URG data pointerchecksum
FSRPAUheadlen
notused
options (variable length)
URG urgent data (generally not used)
ACK ACK valid
PSH push data now(generally not used)
RST SYN FINConnection mgmt
(setup teardowncommands)
bytes rcvr willingto accept
countingby bytes of data(not segments)
Internetchecksum
(as in UDP)
32-bit wordsin header
Transport Layer 3-7
TCP seq numbers ACKssequence numbers
byte stream ldquonumberrdquo of first byte in segmentrsquos data
acknowledgementssequence of next byte expected from other side
cumulative ACKQ how receiver handles out-of-order segmentsA TCP spec doesnrsquot say - up to implementer
A SACK option possible per RFC 2018
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
incoming segment to sender
A
sent ACKed
sent not-yet ACKed(ldquoin-flightrdquo)
usablebut not yet sent
not usable
window size N
sender sequence number space
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
outgoing segment from sender
Transport Layer 3-8
TCP seq numbers ACKs
Usertypes
lsquoCrsquo
host ACKsreceipt
of echoedlsquoCrsquo
host ACKsreceipt oflsquoCrsquo echoesback lsquoCrsquo
simple telnet scenario
Host BHost A
Seq=42 ACK=79 data = lsquoCrsquo
Seq=79 ACK=43 data = lsquoCrsquo
Seq=43 ACK=80
Transport Layer 3-9
TCP round trip time timeoutQ how to set TCP
timeout value longer than RTT
but RTT varies too short premature
timeout unnecessary retransmissions
too long slow reaction to segment loss
Q how to estimate RTT
SampleRTT measured time from segment transmission until ACK receipt ldquobest practicerdquo uses TCP
timer option per RFC 1323 ignore retransmissions
SampleRTT will vary so we want estimated RTT to be ldquosmootherrdquo average several recent
measurements not just current SampleRTT
Transport Layer 3-10
RTT gaiacsumassedu to fantasiaeurecomfr
100
150
200
250
300
350
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
RTT
(mill
isec
onds
)
SampleRTT Estimated RTT
EstimatedRTT = (1- )EstimatedRTT + SampleRTT exponential weighted moving average influence of past sample decreases
exponentially fast typical value = 0125
TCP round trip time timeout
RTT
(mill
iseco
nds)
RTT gaiacsumassedu to fantasiaeurecomfr
sampleRTT
EstimatedRTT
time (seconds)
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
TCP Logical End-to-End Connection
Transport Layer 3-5
socketdoor
TCP send buffer
TCP receive buffer
socketdoor
segment
application processwrites data
application processwrites data
segment
a TCP connection is point-to-point only between a single sender and a single receiver
Multicast with TCP is not possible
a TCP connection is point-to-point only between a single sender and a single receiver
Multicast with TCP is not possible
Transport Layer 3-6
TCP segment structure
source port dest port
32 bits
applicationdata
(variable length)
sequence number
acknowledgement number
receive window
URG data pointerchecksum
FSRPAUheadlen
notused
options (variable length)
URG urgent data (generally not used)
ACK ACK valid
PSH push data now(generally not used)
RST SYN FINConnection mgmt
(setup teardowncommands)
bytes rcvr willingto accept
countingby bytes of data(not segments)
Internetchecksum
(as in UDP)
32-bit wordsin header
Transport Layer 3-7
TCP seq numbers ACKssequence numbers
byte stream ldquonumberrdquo of first byte in segmentrsquos data
acknowledgementssequence of next byte expected from other side
cumulative ACKQ how receiver handles out-of-order segmentsA TCP spec doesnrsquot say - up to implementer
A SACK option possible per RFC 2018
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
incoming segment to sender
A
sent ACKed
sent not-yet ACKed(ldquoin-flightrdquo)
usablebut not yet sent
not usable
window size N
sender sequence number space
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
outgoing segment from sender
Transport Layer 3-8
TCP seq numbers ACKs
Usertypes
lsquoCrsquo
host ACKsreceipt
of echoedlsquoCrsquo
host ACKsreceipt oflsquoCrsquo echoesback lsquoCrsquo
simple telnet scenario
Host BHost A
Seq=42 ACK=79 data = lsquoCrsquo
Seq=79 ACK=43 data = lsquoCrsquo
Seq=43 ACK=80
Transport Layer 3-9
TCP round trip time timeoutQ how to set TCP
timeout value longer than RTT
but RTT varies too short premature
timeout unnecessary retransmissions
too long slow reaction to segment loss
Q how to estimate RTT
SampleRTT measured time from segment transmission until ACK receipt ldquobest practicerdquo uses TCP
timer option per RFC 1323 ignore retransmissions
SampleRTT will vary so we want estimated RTT to be ldquosmootherrdquo average several recent
measurements not just current SampleRTT
Transport Layer 3-10
RTT gaiacsumassedu to fantasiaeurecomfr
100
150
200
250
300
350
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
RTT
(mill
isec
onds
)
SampleRTT Estimated RTT
EstimatedRTT = (1- )EstimatedRTT + SampleRTT exponential weighted moving average influence of past sample decreases
exponentially fast typical value = 0125
TCP round trip time timeout
RTT
(mill
iseco
nds)
RTT gaiacsumassedu to fantasiaeurecomfr
sampleRTT
EstimatedRTT
time (seconds)
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-6
TCP segment structure
source port dest port
32 bits
applicationdata
(variable length)
sequence number
acknowledgement number
receive window
URG data pointerchecksum
FSRPAUheadlen
notused
options (variable length)
URG urgent data (generally not used)
ACK ACK valid
PSH push data now(generally not used)
RST SYN FINConnection mgmt
(setup teardowncommands)
bytes rcvr willingto accept
countingby bytes of data(not segments)
Internetchecksum
(as in UDP)
32-bit wordsin header
Transport Layer 3-7
TCP seq numbers ACKssequence numbers
byte stream ldquonumberrdquo of first byte in segmentrsquos data
acknowledgementssequence of next byte expected from other side
cumulative ACKQ how receiver handles out-of-order segmentsA TCP spec doesnrsquot say - up to implementer
A SACK option possible per RFC 2018
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
incoming segment to sender
A
sent ACKed
sent not-yet ACKed(ldquoin-flightrdquo)
usablebut not yet sent
not usable
window size N
sender sequence number space
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
outgoing segment from sender
Transport Layer 3-8
TCP seq numbers ACKs
Usertypes
lsquoCrsquo
host ACKsreceipt
of echoedlsquoCrsquo
host ACKsreceipt oflsquoCrsquo echoesback lsquoCrsquo
simple telnet scenario
Host BHost A
Seq=42 ACK=79 data = lsquoCrsquo
Seq=79 ACK=43 data = lsquoCrsquo
Seq=43 ACK=80
Transport Layer 3-9
TCP round trip time timeoutQ how to set TCP
timeout value longer than RTT
but RTT varies too short premature
timeout unnecessary retransmissions
too long slow reaction to segment loss
Q how to estimate RTT
SampleRTT measured time from segment transmission until ACK receipt ldquobest practicerdquo uses TCP
timer option per RFC 1323 ignore retransmissions
SampleRTT will vary so we want estimated RTT to be ldquosmootherrdquo average several recent
measurements not just current SampleRTT
Transport Layer 3-10
RTT gaiacsumassedu to fantasiaeurecomfr
100
150
200
250
300
350
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
RTT
(mill
isec
onds
)
SampleRTT Estimated RTT
EstimatedRTT = (1- )EstimatedRTT + SampleRTT exponential weighted moving average influence of past sample decreases
exponentially fast typical value = 0125
TCP round trip time timeout
RTT
(mill
iseco
nds)
RTT gaiacsumassedu to fantasiaeurecomfr
sampleRTT
EstimatedRTT
time (seconds)
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-7
TCP seq numbers ACKssequence numbers
byte stream ldquonumberrdquo of first byte in segmentrsquos data
acknowledgementssequence of next byte expected from other side
cumulative ACKQ how receiver handles out-of-order segmentsA TCP spec doesnrsquot say - up to implementer
A SACK option possible per RFC 2018
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
incoming segment to sender
A
sent ACKed
sent not-yet ACKed(ldquoin-flightrdquo)
usablebut not yet sent
not usable
window size N
sender sequence number space
source port dest port
sequence number
acknowledgement number
checksum
rwnd
urg pointer
outgoing segment from sender
Transport Layer 3-8
TCP seq numbers ACKs
Usertypes
lsquoCrsquo
host ACKsreceipt
of echoedlsquoCrsquo
host ACKsreceipt oflsquoCrsquo echoesback lsquoCrsquo
simple telnet scenario
Host BHost A
Seq=42 ACK=79 data = lsquoCrsquo
Seq=79 ACK=43 data = lsquoCrsquo
Seq=43 ACK=80
Transport Layer 3-9
TCP round trip time timeoutQ how to set TCP
timeout value longer than RTT
but RTT varies too short premature
timeout unnecessary retransmissions
too long slow reaction to segment loss
Q how to estimate RTT
SampleRTT measured time from segment transmission until ACK receipt ldquobest practicerdquo uses TCP
timer option per RFC 1323 ignore retransmissions
SampleRTT will vary so we want estimated RTT to be ldquosmootherrdquo average several recent
measurements not just current SampleRTT
Transport Layer 3-10
RTT gaiacsumassedu to fantasiaeurecomfr
100
150
200
250
300
350
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
RTT
(mill
isec
onds
)
SampleRTT Estimated RTT
EstimatedRTT = (1- )EstimatedRTT + SampleRTT exponential weighted moving average influence of past sample decreases
exponentially fast typical value = 0125
TCP round trip time timeout
RTT
(mill
iseco
nds)
RTT gaiacsumassedu to fantasiaeurecomfr
sampleRTT
EstimatedRTT
time (seconds)
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-8
TCP seq numbers ACKs
Usertypes
lsquoCrsquo
host ACKsreceipt
of echoedlsquoCrsquo
host ACKsreceipt oflsquoCrsquo echoesback lsquoCrsquo
simple telnet scenario
Host BHost A
Seq=42 ACK=79 data = lsquoCrsquo
Seq=79 ACK=43 data = lsquoCrsquo
Seq=43 ACK=80
Transport Layer 3-9
TCP round trip time timeoutQ how to set TCP
timeout value longer than RTT
but RTT varies too short premature
timeout unnecessary retransmissions
too long slow reaction to segment loss
Q how to estimate RTT
SampleRTT measured time from segment transmission until ACK receipt ldquobest practicerdquo uses TCP
timer option per RFC 1323 ignore retransmissions
SampleRTT will vary so we want estimated RTT to be ldquosmootherrdquo average several recent
measurements not just current SampleRTT
Transport Layer 3-10
RTT gaiacsumassedu to fantasiaeurecomfr
100
150
200
250
300
350
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
RTT
(mill
isec
onds
)
SampleRTT Estimated RTT
EstimatedRTT = (1- )EstimatedRTT + SampleRTT exponential weighted moving average influence of past sample decreases
exponentially fast typical value = 0125
TCP round trip time timeout
RTT
(mill
iseco
nds)
RTT gaiacsumassedu to fantasiaeurecomfr
sampleRTT
EstimatedRTT
time (seconds)
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-9
TCP round trip time timeoutQ how to set TCP
timeout value longer than RTT
but RTT varies too short premature
timeout unnecessary retransmissions
too long slow reaction to segment loss
Q how to estimate RTT
SampleRTT measured time from segment transmission until ACK receipt ldquobest practicerdquo uses TCP
timer option per RFC 1323 ignore retransmissions
SampleRTT will vary so we want estimated RTT to be ldquosmootherrdquo average several recent
measurements not just current SampleRTT
Transport Layer 3-10
RTT gaiacsumassedu to fantasiaeurecomfr
100
150
200
250
300
350
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
RTT
(mill
isec
onds
)
SampleRTT Estimated RTT
EstimatedRTT = (1- )EstimatedRTT + SampleRTT exponential weighted moving average influence of past sample decreases
exponentially fast typical value = 0125
TCP round trip time timeout
RTT
(mill
iseco
nds)
RTT gaiacsumassedu to fantasiaeurecomfr
sampleRTT
EstimatedRTT
time (seconds)
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-10
RTT gaiacsumassedu to fantasiaeurecomfr
100
150
200
250
300
350
1 8 15 22 29 36 43 50 57 64 71 78 85 92 99 106
time (seconnds)
RTT
(mill
isec
onds
)
SampleRTT Estimated RTT
EstimatedRTT = (1- )EstimatedRTT + SampleRTT exponential weighted moving average influence of past sample decreases
exponentially fast typical value = 0125
TCP round trip time timeout
RTT
(mill
iseco
nds)
RTT gaiacsumassedu to fantasiaeurecomfr
sampleRTT
EstimatedRTT
time (seconds)
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-11
timeout interval EstimatedRTT plus ldquosafety marginrdquo large variation in EstimatedRTT -gt larger safety margin
estimate SampleRTT deviation from EstimatedRTT
DevRTT = (1-)DevRTT + |SampleRTT-EstimatedRTT|
TCP round trip time timeout
(typically = 025)
TimeoutInterval(RTO) = EstimatedRTT + 4DevRTT
estimated RTT ldquosafety marginrdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-12
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-13
TCP reliable data transfer TCP creates rdt
service on top of IPrsquos unreliable service pipelined segments cumulative acks single
retransmission timer retransmissions
triggered by timeout events duplicate acks
letrsquos initially consider simplified TCP sender ignore duplicate acks ignore flow control
congestion control
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-14
TCP sender eventsdata rcvd from app create segment with
seq seq is byte-
stream number of first data byte in segment
start timer if not already running think of timer as for
oldest unacked segment
expiration interval TimeOutInterval
timeout retransmit segment
that caused timeout restart timer ack rcvd if ack acknowledges
previously unacked segments update what is
known to be ACKed start timer if there
are still unacked segments
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-15
TCP sender (simplified)
waitfor
event
NextSeqNum = InitialSeqNumSendBase = InitialSeqNum
create segment seq NextSeqNumpass segment to IP (ie ldquosendrdquo)NextSeqNum = NextSeqNum + length(data) if (timer currently not running) start timer
data received from application above
retransmit not-yet-ACKed segment with smallest seq restart timer
timeout
if (y gt SendBase) SendBase = y SendBasendash1 last cumulatively ACKed byte if (there are currently not-yet-ACKed segments) restart timer else stop timer
ACK received with ACK field value y
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-16
TCP retransmission scenarios
lost ACK scenario
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8 bytes of data
Xtim
eo
ut
ACK=100
premature timeout
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=92 8bytes of data
tim
eo
ut
ACK=120
Seq=100 20 bytes of data
ACK=120
SendBase=100
SendBase=120
SendBase=120
SendBase=92
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-17
TCP retransmission scenarios
X
cumulative ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
Seq=120 15 bytes of data
tim
eo
ut
Seq=100 20 bytes of data
ACK=120
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-18
TCP ACK generation [RFC 1122 RFC
2581 5681]
event at receiver
arrival of in-order segment withexpected seq All data up toexpected seq already ACKed
arrival of in-order segment withexpected seq One other segment has ACK pending
arrival of out-of-order segmenthigher-than-expect seq Gap detected
arrival of segment that partially or completely fills gap
TCP receiver action
delayed ACK Wait up to 500msfor next segment If no next segmentsend ACK
immediately send single cumulative ACK ACKing both in-order segments
immediately send duplicate ACK indicating seq of next expected byte
immediate send ACK provided thatsegment starts at lower end of gap
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-19
TCP fast retransmit time-out period
often relatively long long delay before
resending lost packet detect lost
segments via duplicate ACKs sender often sends
many segments back-to-back
if segment is lost there will likely be many duplicate ACKs
if sender receives 3 ACKs for same data(ldquotriple duplicate ACKsrdquo) resend unACKed segment with smallest sequence
likely that unacked segment lost so donrsquot wait for timeout
TCP fast retransmit
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-20
X
fast retransmit after sender receipt of triple duplicate ACK
Host BHost A
Seq=92 8 bytes of data
ACK=100
tim
eo
ut ACK=100
ACK=100
ACK=100
TCP fast retransmit
Seq=100 20 bytes of data
Seq=100 20 bytes of data
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-21
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-22
TCP flow controlapplication
process
TCP socketreceiver buffers
TCPcode
IPcode
application
OS
receiver protocol stack
receiverrsquos application may remove data from
TCP socket buffer hellip
hellip slower than TCP is delivering
it to the buffer
(sender is sending)
from sender
receiver controls sender so sender wonrsquot overflow receiverrsquos buffer by transmitting too much too fast
flow control
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-23
TCP flow control
buffered data
free buffer spacerwnd
RcvBuffer
TCP segment payloads
to application process
receiver ldquoadvertisesrdquo free buffer space by including rwnd value in TCP header of receiver-to-sender segments RcvBuffer size is set by
operating system via socket options (typical default is 4096 bytes)
many operating systems autoadjust RcvBuffer based on available resources
sender limits amount of unACKed (ldquoin-flightrdquo) data to receiverrsquos rwnd value
guarantees receive buffer will not overflow
receiver-side buffering
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-24
TCP flow control receiver OS tracks
rwnd current size of its receive window LastByteReceived bytestream number of last byte placed in
buffer LastByteRead bytestream number of last byte read from
buffer
hellipand informs sender of its available buffer space by setting TCP header field in itrsquos acknowledgment segments as
rwnd = RcvBuffer ndash [LastByteReceived ndash LastByteRead]
sender OS tracks LastByteSent bytestream number of last byte sent to receiver LastByteACKed bytestream number of last byte acknowledged
by receiver
hellipand restricts sending rate such thatLastByteSent ndash LastByteACKed rwnd
Q What happens if receive buffer becomes full so that rwnd = 0
rwnd = 4096 ndash [120000 ndash 118000] = 4096 - 2000 = 2096
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-25
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-26
Connection Managementbefore exchanging data sender amp receiver
ldquohandshakerdquo agree to establish connection (each knowing the
other willing to establish connection) agree on connection parameters
connection state ESTABconnection variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
connection state ESTABconnection Variables
seq client-to-server server-to-clientrcvBuffer size at serverclient
application
network
Socket clientSocket = newSocket(hostnameport
number)
Socket connectionSocket = welcomeSocketaccept()
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-27
Q will 2-way handshake always work in network
variable delays retransmitted messages
(eg req_conn(x)) due to message loss
message reordering canrsquot ldquoseerdquo other side
2-way handshake
Letrsquos talk
OKESTAB
ESTAB
choose xreq_conn(x)
ESTAB
ESTABacc_conn(x)
Agreeing to establish a connection
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-28
Agreeing to establish a connection
2-way handshake failure scenarios
retransmitreq_conn(
x)
ESTAB
req_conn(x)
half open connection(no client)
client terminat
es
serverforgets x
connection x completes
retransmitreq_conn(
x)
ESTAB
req_conn(x)
data(x+1)
retransmitdata(x+1)
acceptdata(x+1)
choose xreq_conn(x)
ESTAB
ESTAB
acc_conn(x)
client terminat
es
ESTAB
choose xreq_conn(x)
ESTAB
acc_conn(x)
data(x+1) acceptdata(x+1)
connection x completes server
forgets x
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-29
TCP 3-way handshake
SYNbit=1 Seq=x
choose init seq num xsend TCP SYN msg
ESTAB
SYNbit=1 Seq=yACKbit=1 ACKnum=x+1
choose init seq num ysend TCP SYNACKmsg acking SYN
ACKbit=1 ACKnum=y+1
received SYNACK(x) indicates server is livesend ACK for SYNACK
this segment may contain client-to-server data
received ACK(y) indicates client is live
SYNSENT
ESTAB
SYN RCVD
client state
LISTEN
server state
LISTEN
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-30
TCP 3-way handshake FSM
closed
listen
SYNrcvd
SYNsent
ESTAB
Socket clientSocket = newSocket(hostnameport
number)
SYN(seq=x)
Socket connectionSocket = welcomeSocketaccept()
SYN(x)
SYNACK(seq=yACKnum=x+1)create new socket for
communication back to client
SYNACK(seq=yACKnum=x+1)
ACK(ACKnum=y+1)ACK(ACKnum=y+1)
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-31
TCP closing a connection client server each close their side of
connection send TCP segment with FIN bit = 1
respond to received FIN with ACK on receiving FIN ACK can be combined with
own FIN simultaneous FIN exchanges can be
handled
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-32
FIN_WAIT_2
CLOSE_WAIT
FINbit=1 seq=y
ACKbit=1 ACKnum=y+1
ACKbit=1 ACKnum=x+1 wait for server
close
can stillsend data
can no longersend data
LAST_ACK
CLOSED
TIMED_WAIT
timed wait for 2max
segment lifetime
CLOSED
TCP closing a connection
FIN_WAIT_1 FINbit=1 seq=xcan no longersend but can receive data
clientSocketclose()
client state server state
ESTABESTAB
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-33
TCP connection life cycle
TCP clientlifecycle
TCP serverlifecycle
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-34
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-35
congestion informally ldquotoo many sources sending sending
too much too much data too fast too fast for network to handlerdquo
different from flow control manifestations
lost packets (buffer overflow at routers)
long delays (queuing in router buffers) another top-10 problem
Principles of congestion control
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-36
Causescosts of congestion scenario 1
two senders two receivers
Host apps generates data at rate in
one router infinite buffers
output link capacity R no retransmission
flow control etc
maximum per-connection throughput R2
unlimited shared output link buffers
Host A
original data in
Host B
throughputout
R2
R2
out
in R2d
ela
yin
large delays as arrival rate in approaches capacity
R
Recall traffic
intensity
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-37
one router finite buffers sender retransmission of timed-out packet
application-layer input = application-layer outputin
= out
transport-layer input includes retransmissions in in
finite shared output link buffers
Host A
in original data
Host B
outin original data plus
retransmitted data
lsquo
Causescosts of congestion scenario 2
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-38
idealization perfect knowledge
sender sends only when router buffers available
finite shared output link buffers
in original dataoutin original data plus
retransmitted data
copy
free buffer space
R2
R2
out
in
Causescosts of congestion scenario 2
Host B
A
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-39
in original dataoutin original data plus
retransmitted data
copy
no buffer space
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
Causescosts of congestion scenario 2
A
Host B
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-40
in original dataoutin original data plus
retransmitted data
free buffer space
Causescosts of congestion scenario 2
Idealization known loss packets can be lost dropped at router due to full buffers
sender only resends if packet known to be lost
R2
R2in
out
when sending at R2 some packets are retransmissions but asymptotic goodput is still R2 (why)
A
Host B
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-41
A
in outincopy
free buffer space
timeout
R2
R2in
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
Host B
Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Causescosts of congestion scenario 2
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-42
R2
out
when sending at R2 some packets are retransmissions including duplicates that are delivered
ldquocostsrdquo of congestion more work (retrans) to compensate for lost
packets unneeded retransmissions link carries multiple
copies of packet
R2in
Causescosts of congestion scenario 2 Realistic duplicates packets can be lost
dropped at router due to full buffers
sender times out prematurely sending two copies both of which are delivered
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-43
four senders multihop paths timeoutretransmit
Q what happens as in and in
rsquo increase
finite shared output link buffers
Host A out
Causescosts of congestion scenario 3
Host B
Host C
Host D
in original data
in original data plus
retransmitted data
A as red inrsquo increases all
arriving blue pkts at upper queue are dropped blue throughput 0
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-44
another ldquocostrdquo of congestion when packet dropped any ldquoupstreamrdquo
transmission capacity used for that packet was wasted
Causescosts of congestion scenario 3
C2
C2
ou
t
inrsquo
bullbuffers fill toward capacitybullpackets discardeddelayedbullsources re-transmit lost
packetsbullgood packets are resent
(ack lostdelayed)bull routers generate more
traffic to update pathsbullDelaysloads propagate
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-45
Approaches towards congestion controltwo broad approaches towards congestion
controlend-end
congestion control
no explicit feedback from network
congestion inferred from end-system observed loss delay
approach taken by TCP
network-assisted congestion control
routers provide feedback to end systemssingle bit indicating congestion (SNA DECbit TCPIP ECN ATM)
explicit send rate for sender
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-46
Case study ATM ABR congestion control
ABR available bit rate
ldquoelastic servicerdquo if senderrsquos path
ldquounderloadedrdquo sender should
use available bandwidth
if senderrsquos path congested sender throttled
to minimum guaranteed rate
RM (resource management) cells
sent by sender interspersed with data cells
bits in RM cell set by switches (ldquonetwork-assistedrdquo) NI bit no increase in rate
(mild congestion) CI bit congestion
indication RM cells returned to sender
by receiver with bits intact
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-47
Case study ATM ABR congestion control
two-byte ER (explicit rate) field in RM cell congested switch may lower ER value in cell sendersrsquo send rate thus max supportable rate on path
EFCI bit in data cells set to 1 in congested switch if data cell preceding RM cell has EFCI set receiver
sets CI bit in returned RM cell
RM cell data cell
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-48
Chapter 3 outline
31 transport-layer services
32 multiplexing and demultiplexing
33 connectionless transport UDP
34 principles of reliable data transfer
35 connection-oriented transport TCP segment structure reliable data transfer flow control connection
management36 principles of
congestion control37 TCP congestion
control
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-49
TCP congestion control additive increase multiplicative decrease
approach sender increases transmission rate (window size) probing for usable bandwidth until loss occurs additive increase increase cwnd by 1
MSS every RTT until loss detected multiplicative decrease cut cwnd in half
after loss
cwnd
TC
P s
ende
r co
nges
tion
win
dow
siz
e
AIMD saw toothbehavior probing
for bandwidth
additively increase window size helliphellip until loss occurs (then cut window in half)
time
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-50
TCP Congestion Control details
sender limits transmission
cwnd is dynamic and a function of perceived network congestion
TCP sending rate roughly send
cwnd bytes wait RTT for ACKS then send more bytes
last byteACKed sent not-yet
ACKed(ldquoin-flightrdquo)
last byte sent
cwndsender sequence number space
rate ~~cwnd
RTTbytessec
LastByteSent-LastByteAcked
lt mincwndrwnd
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-51
TCP Slow Start when connection
begins increase rate exponentially until first loss event initially cwnd = 1 MSS increment cwnd by 1
MSS for every ACK received
effect is doubling of cwnd size every RTT
result initial rate is slow but ramps up exponentially fast
Host A
one segment
RT
T
Host B
time
two segments
four segments
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-52
TCP detecting reacting to loss
loss indicated by timeout cwnd set to 1 MSS window then grows exponentially (as in slow start) to threshold then
grows linearly loss indicated by 3 duplicate ACKs TCP RENO
dup ACKs indicate network capable of delivering some segments cwnd is cut in half (+3 MSS) window then grows linearly
TCP Tahoe always sets cwnd to 1 (timeout or 3 duplicate acks) then slowstart
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-53
Q when should the exponential increase switch to linear
A when cwnd gets to 12 of its value before timeout
Implementation variable ssthresh on loss event ssthresh is set to 12 of cwnd just before loss event
TCP switching from slow start to CA
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-54
Summary TCP Congestion Control
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd gt ssthresh
congestionavoidance
cwnd = cwnd + MSS (MSScwnd)dupACKcount = 0
transmit new segment(s) as allowed
new ACK
dupACKcount++
duplicate ACK
fastrecovery
cwnd = cwnd + MSStransmit new segment(s) as allowed
duplicate ACK
ssthresh= cwnd2cwnd = ssthresh + 3
retransmit missing segment
dupACKcount == 3
timeoutssthresh = cwnd2cwnd = 1 dupACKcount = 0retransmit missing segment
ssthresh= cwnd2cwnd = ssthresh + 3retransmit missing segment
dupACKcount == 3cwnd = ssthreshdupACKcount = 0
New ACK
slow start
timeoutssthresh = cwnd2
cwnd = 1 MSSdupACKcount = 0
retransmit missing segment
cwnd = cwnd+MSSdupACKcount = 0transmit new segment(s) as allowed
new ACKdupACKcount++
duplicate ACK
cwnd = 1 MSS
ssthresh = 64 KBdupACKcount = 0
NewACK
NewACK
NewACK
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-55
TCP throughput avg TCP thruput as function of window
size RTT ignore slow start assume always data to send
W window size (measured in bytes) where loss occurs avg window size ( in-flight bytes) is frac34 W avg thruput is 34W per RTT
W
W2
avg TCP thruput = 34W
RTTbytessec
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-56
TCP Futures TCP over ldquolong fat pipesrdquo example 1500 byte segments 100ms RTT
want 10 Gbps throughput requires W = 83333 in-flight segments throughput in terms of segment loss
probability L [Mathis 1997]
to achieve 10 Gbps throughput need a loss rate of L = 210-10 or one loss event every 5000000000 segments ndash a very small loss rate
new versions of TCP for high-speed
TCP throughput = 122 MSSRTT L
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-57
fairness goal if K TCP sessions share same bottleneck link of bandwidth R each should have average rate of RK
TCP connection 1
bottleneckrouter
capacity R
TCP Fairness
TCP connection 2
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-58
Why is TCP fairtwo competing sessions additive increase gives slope of 1 as throughout
increases multiplicative decrease decreases throughput
proportionally R
R
equal bandwidth share
Connection 1 throughput
Con
nect
ion
2 th
roug
h pu t
congestion avoidance additive increaseloss decrease window by factor of 2
congestion avoidance additive increaseloss decrease window by factor of 2
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-59
Fairness (more)Fairness and UDP multimedia apps
often do not use TCP do not want rate
throttled by congestion control
instead use UDP send audiovideo
at constant rate tolerate packet loss
Fairness parallel TCP connections
application can open multiple parallel connections between two hosts
web browsers do this eg link of rate R with 9
existing connections new app asks for 1 TCP gets
rate R10 new app asks for 11 TCPs gets
R2
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
Transport Layer 3-60
Chapter 3 summary principles behind transport
layer services multiplexing
demultiplexing reliable data transfer flow control congestion control
instantiation implementation in the Internet UDP TCP
next leaving the
network ldquoedgerdquo (application transport layers)
into the network ldquocorerdquo
