Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Peer-to-Peer Networks01: Organization and Introduction
Christian SchindelhauerTechnical FacultyComputer-Networks and TelematicsUniversity of Freiburg
Web & Dates
Web page- http://cone.informatik.uni-freiburg.de/cone_teach/
cone_teach_current/p2p-WS12
Lecture- Monday, 10am-12pm, 106-00-007- Wednesday, 10am-11am, 106-00-007
Exercise classes- Christian Ortolf- Wednesday, 11am-12pm, building 101, 106-00-007
2
Exercises
Exercise class- Wednesday, 11am-12pm, building 101, 106-00-007 - start: 31.11.2012
Exercises- appear every Wednesay on the web-page- should be solved by students- are the basis for the oral exam- solutions of the exercises are discussed in the following
week
3
Exam
Oral exam- based on the lecture and the exercises- register online for the exam- Mandatory registration
4
Materials
Slides- appear before the lecture on the web-
page
Book- at least 70% of the lecture can be
found in Mahlmann, Schindelhauer, Peer-to-Peer-Netzwerke — Methoden und Algorithmen, Springer 2007
Further Literature- Research papers will be presented
during the lecture on the slides and on the web-page
5
White Paper
© 2011 Cisco and/or its affiliates. All rights reserved. This document is Cisco Public Information. Page 5 of 27
Table 3. Global Mobile Data Growth Today is Similar to Global Internet Growth (Fixed) in the Late 1990s
Global Internet Traffic Growth Global Mobile Data Traffic Growth
1997 178% 2008 156%
1998 124% 2009 140%
1999 128% 2010 159%
2000 195% 2011 (estimate) 131%
2001 133% 2012 (estimate) 113%
Source: Cisco VNI Mobile, 2011
In the long term, mobile data and fixed traffic should settle into the same growth rate, although the mobile data growth rate is likely to remain higher than the fixed growth rate for the next 7 to 10 years.
Global Mobile Data Traffic, 2010 to 2015 Overall mobile data traffic is expected to grow to 6.3 exabytes per month by 2015, a 26-fold increase over 2010. Mobile data traffic will grow at a CAGR of 92 percent from 2010 to 2015. Annual growth rates will taper over the forecast period from 131 percent in 2011 to 64 percent in 2015 (Figure 1).
Figure 1. Cisco Forecasts 6.3 Exabytes per Month of Mobile Data Traffic by 2015
Western Europe and Asia Pacific will account for over half of global mobile traffic by 2015, as shown in Figure 2. Middle East and Africa will experience the highest CAGR of 129 percent, increasing 63-fold over the forecast period. The emerging market regions (Central and Eastern Europe, Latin America, and Middle East and Africa) will have the highest growth and will represent an increasing share of total mobile data traffic, from 12 percent at the end of 2010 to 20 percent by 2015.
Internet Traffic6
7
Increase of Internet Traffic
20102005200019951990 2015
1 Mbit/s
1 Gbit/s
1 Tbit/s
10 Mbit/s
100 Mbit/s
10 Gbit/s
100 Gbit/s
10 Tbit/s
100 Tbit/s
1 Pbit/s
Cisco: 966 Exabyte/y 2015
2020
10 Pbit/s
100 Pbit/s
1 Ebit/s
Internet-Verkehr
Global Internet Traffic Shares1993-2004
Source: CacheLogic 2005
8
FTP
Peer-to-Peer
Web
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
CacheLogic Research Trends of Internet Protocols 1993-2004
Share
of In
tern
et tr
affi
c
0
10
20
30
40
50
60
70
P2P and Filesharing Traffic
Source: Ipoque 2007
9
Source: Ipoque 2007
P2P Systems Germany 2007by Volume
Source: Ipoque 2007
10
What Germans Download 2007by Volume
Source: Ipoque 2007
11
Global Internet Traffic 2007
Ellacoya report (June 2007)- worldwide
HTTP traffic volume overtakes P2P after four years continues record
Main reason: Youtube.com
12
rest2 %
VoIP1 %
Newsgroups9 %
non-HTTP video streaming3 %
gaming2 %
P2P37 %
HTTP46 %
Internet Traffic 2010Motivation Related Work Data Set BitTorrent Analysis Conclusion
Related Work: Internet Analysis
Cisco Visual NetworkingIndex Usage
contains data of 20anonymous service providers
Tra�c Study
Traf
fic [P
erce
ntag
e]
0
10
20
30
40
50
HTTP
Online V
ideo
P2P File
Sharin
g
Web−B
ased
FS Rest
26.39 26.15 24.85
18.69
3.92
Filesharing43.54
[”Cisco Visual Networking Index: Usage”, White Paper, 2010]
Thomas Janson (University of Freiburg) Analysis of P2P Tra�c and User Behaviour September 7th, 2011 3 / 15
HTTP 44.4 %
BitTorrent 24.1 %
NNTP 14.2 %SHOUTcast 6.4 %
RTMP 5 %
eDonkey 4 %RTSP 1.2 %Skype 0.8 %
HTTP 14.6 %BitTorrent 64.3 %
NNTP 0.7 %SHOUTcast 0.7 %RTMP 0.4 %
eDonkey 16.3 %
RTSP 0.1 %Skype 3 %
Internet Traffic of a German ISPAugust 2009
14
Download
Upload
Source: Alsbih, Janson, S. Analysis of Peer-to-Peer Traffic and User Behaviour ITA 2011
Internet Traffic of a German ISPAugust 2009
15
Motivation Related Work Data Set BitTorrent Analysis Conclusion
Services
HTTP most tra�c BitTorrent most upload
Top ten services of the average user
Mea
n Ho
st T
raffi
c [k
b/s]
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
HTTP in
HTTP out
BitTorre
nt in
BitTorre
nt ou
t
NNTP in
NNTP out
eDon
key in
eDon
key o
utSSL
in
SSL ou
t
SHOUTcast
in
SHOUTcast
out
RTMP in
RTMP out
FTP trans
fer in
FTP trans
fer o
ut
Micros
oft BITS in
Micros
oft BITS o
ut
Gnutel
la in
Gnutel
la ou
t
Rest i
n
Rest o
ut
3.25
0.12
0.680.52
0.34
0.010.210.24 0.21
0.03 0.140.02 0.11 0 0.060.01 0.05 0 0.04 0
0.390.21
!""#$%&'$
()*"+,,-.*$/&'$
0-1*$23'$
!""#$%%&$
'()"*++,-)$./&$
0,1)$..&$
Download Upload
5x
4x
Thomas Janson (University of Freiburg) Analysis of P2P Tra�c and User Behaviour September 7th, 2011 6 / 15
Source: Alsbih, Janson, S. Analysis of Peer-to-Peer Traffic and User Behaviour ITA 2011
BitTorrent User Behavior of a German ISPAugust 2009
16
Motivation Related Work Data Set BitTorrent Analysis Conclusion
BitTorrent User Behaviour
online period lengthis a piecewise powerlaw distribution
interval bounds arethe 16th and 24thhour ) user let theclient run overnight or not
Online period length probability●
●
●●
●●
●●
● ●● ● ● ● ●●
●●●●●●
●●●●●
●●●●●●●●●●●
●●
●●
●●●
●●
●●●●●
●●
●
●●●●●
●
●●●●
●
●
●●●●●●●●
●●
●
●●
●
●●
●●
●
●
●
●
●
●●●
●
●
●
●
●●●
●
●
●
●
●●●●●
●
●●
●
●
●
●
●
●●
●●
●●
●
●●
●●
●●
●●
●
●
●
●●●●●●●●
●
●●
●
●
●
●
●●●
1 2 5 10 20 50 100
5e−0
51e−0
35e−0
2
Probability Distribution Online Period
online period [hours][log]
prob
abilit
y [lo
g]
16 24
probability for online period length [in hours]approximated function
cases of piecewise definition
P [online period t] ⇡
8><
>:
0.18 · t�0.82 for t � 16, (� = 0.013)
2782 · t�4.40 for 16 < t 24, (� = 0.00006)
11 · t�2.58 for t > 24 (� = 0.000015)Thomas Janson (University of Freiburg) Analysis of P2P Tra�c and User Behaviour September 7th, 2011 11 / 15
Source: Alsbih, Janson, S. Analysis of Peer-to-Peer Traffic and User Behaviour ITA 2011
BitTorrent User Behavior of a German ISPAugust 2009
17
Motivation Related Work Data Set BitTorrent Analysis Conclusion
BitTorrent Periodicity
Fourier analysis shows 12h and24h peak
24h periodicity roughlyresembles sin curve
Fourier analysis of tra�c & periodicty
0 50
100 150 200 250 300
0 24 48 72 96 120 144 168 192 216 240
ener
gy [M
B/ho
ur]
period [hours]
Incoming TrafficOutgoing Traffic
0 0.5
1 1.5
2 2.5
3 3.5
4
0 1 2 3 4 5 6 7 8 9 10 11
traffi
c [k
b/s]
half−daytime [hours]
incoming trafficoutgoing traffic
0 0.5
1 1.5
2 2.5
3 3.5
4
0 24 48 72 96 120 144tra
ffic
[kb/
s]day
incoming trafficincoming traffic (daily mean)
outgoing trafficoutgoing traffic (daily mean)
Sat Sun Mon Tue Wed Thu Fri 0
0.5 1
1.5 2
2.5 3
3.5 4
0 2 4 6 8 10 12 14 16 18 20 22
traffi
c [k
b/s]
daytime [hours]
incoming trafficoutgoing traffic
Thomas Janson (University of Freiburg) Analysis of P2P Tra�c and User Behaviour September 7th, 2011 13 / 15
Source: Alsbih, Janson, S. Analysis of Peer-to-Peer Traffic and User Behaviour ITA 2011
Milestones P2P Systems
Napster (1st version: 1999-2000) Gnutella (2000), Gnutella-2 (2002) Edonkey (2000)
- later: Overnet usese Kademlia FreeNet (2000)
- Anonymized download
JXTA (2001)- Open source P2P network platform
FastTrack (2001)- known from KaZaa, Morpheus, Grokster
Bittorrent (2001) - only download, no search
Skype (2003)- VoIP (voice over IP), Chat, Video
18
Milestones Theory
Distributed Hash-Tables (DHT) (1997)- introduced for load balancing between web-servers
CAN (2001)- efficient distributed DHT data structure for P2P networks
Chord (2001)- efficient distributed P2P network with logarithmic search time
Pastry/Tapestry (2001)- efficient distributed P2P network using Plaxton routing
Kademlia (2002)- P2P-Lookup based on XOr-Metrik
Many more exciting approaches- Viceroy, Distance-Halving, Koorde, Skip-Net, P-Grid, ...
Recent developments- Network Coding for P2P- Game theory in P2P- Anonymity, Security
19
What is a P2P Network?
What is P2P NOT?- a peer-to-peer network is not a client-server network
Etymology: peer- from latin par = equal- one that is of equal standing with another- P2P, Peer-to-Peer: a relationship between equal partners
Definition- a Peer-to-Peer Network is a communication network between
computers in the Internet• without central control• and without reliable partners
Observation- the Internet can be seen as a large P2P network
20
Contents
Short history First Peer-to-Peer Networks
- Napster- Gnutella
CAN Chord Pastry und Tapestry Game theory P2P traffic Codes P2P in the real world
21
Peer-to-Peer Networks01: Organization and Introduction
Christian SchindelhauerTechnical FacultyComputer-Networks and TelematicsUniversity of Freiburg