Comparison of Routing Metrics for Static Multi-Hop Wireless Networks

Comparison of Routing Metrics for Static Multi-Hop Wireless Networks

Richard Draves, Jitendra Padhye and Brian ZillMicrosoft Research

Presented by Hoang Nguyen

CS598JH - Spring 06

Some slides adopted from the authors and from Professor Robin Kravets

Multi-hop Wireless Networks

Static Mobile

Motivating scenario

Community wireless

networks (“Mesh Networks”)

Battlefield networks

Key challenge

Improving network capacity

Handling mobility, node failures, limited power.

Routing in Multi-hop Wireless Networks

• Mobile Networks– Minimum-Hop– DSR, AODV…

• Static Networks– Minimum-Hop vs. More Hops

• Link-quality Based Routing– Signal-to-Noise ratio– Packet loss rate– Round-trip-time– Bandwidth

Experimental Comparison

Contributions

• Design and Implementation of LQSR (Link Quality Source Routing) protocol– LQSR = DSR with link quality metrics

• Experimental Comparison of link quality metrics:– Minimum-hop (HOP)– Per-Hop Round Trip Time (RTT)– Per-hop Packet Pair (PktPair)– Expected Transmission (ETX)

Outline

• DSR Revisited

• Link quality metrics revisited

• LQSR (Link Quality Source Routing)

• Experimental results

• Conclusion

Outline

• DSR Revisited




• Conclusion

DSR – Route Discovery

B

A

S EF

H

J

D

CG

IK

Z

Y

M

N

L

Adopted from Professor Robin Kravets’ lectures


Broadcast transmission

[X,Y] Represents list of identifiers appended to RREQ

B

A

S EF

H

J

D

CG

IK

Z

Y

M

N

L

[S]



Node H receives packet RREQ from two neighbors: potential for collision

B

A

S EF

H

J

D

CG

IK

Z

Y

M

N

L

[S,E]

[S,C]



Node C receives RREQ from G and H, but does not forwardit again, because node C has already forwarded RREQ once

B

A

S EF

H

J

D

CG

IK

Z

Y

M

N

L

[S,C,G]

[S,E,F]



Nodes J and K both broadcast RREQ to node DSince nodes J and K are hidden from each other, their transmissions may collide

B

A

S EF

H

J

D

CG

IK

Z

Y

M

N

L

[S,C,G,K]

[S,E,F,J]



Node D does not forward RREQ, because node D is the intended target of the route discovery

B

A

S EF

H

J

D

CG

IK

Z

Y

M

N

L[S,E,F,J,M]


DSR – Route Reply

B

A

S EF

H

J

D

CG

IK

Z

Y

M

N

L

RREP [S,E,F,J,D]


DSR – Data Delivery

Packet header size grows with route length

B

A

S EF

H

J

D

CG

IK

Z

Y

M

N

L

DATA [S,E,F,J,D]


DSR – Route Caching

[P,Q,R] Represents cached route at a node (DSR maintains the cached routes in a tree format)

B

A

S EF

H

J

D

CG

IK

M

N

L

[S,E,F,J,D] [E,F,J,D]

[C,S][G,C,S]

[F,J,D],[F,E,S][J,F,E,S]

Z


DSR – Route Caching

Assume that there is no link between D and Z.Route Reply (RREP) from node K limits flooding of RREQ.In general, the reduction may be less dramatic.

B

A

S EF

H

J

D

CG

IK

Z

Y

M

N

L

[S,E,F,J,D] [E,F,J,D]

[C,S][G,C,S]

[F,J,D],[F,E,S][J,F,E,S]

RREQ[K,G,C,S]

RREP


DSR – Route Error

J sends a route error to S along J-F-E-S when its attempt to forward the data packet S (with route SEFJD) on J-D failsNodes hearing RERR update their route cache to remove link J-D

B

A

S EF

H

J

D

CG

IK

Z

M

N

L

RERR [J-D]


Outline

• DSR Revisited




• Conclusion

Per-hop Round Trip Time (RTT)

• Node periodically pings each of its neighbors

• RTT samples are averaged using exponentially weighted moving average

• Path with least sum of RTTs is selected

Per-hop Round Trip Time (RTT)

• Advantages– Easy to implement– Accounts for link load and bandwidth– Also accounts for link loss rate

• 802.11 retransmits lost packets up to 7 times• Lossy links will have higher RTT

• Disadvantages– Expensive – Self-interference due to queuing

Per-hop Packet Pair (PktPair)

• Node periodically sends two back-to-back probes to each neighbor– First probe is small, second is large

• Neighbor measures delay between the arrival of the two probes; reports back to the sender

• Sender averages delay samples using low-pass filter

• Path with least sum of delays is selected

Per-hop Packet Pair (PktPair)

• Advantages– Self-interference due to queuing is not a

problem– Implicitly takes load, bandwidth and loss rate

into account

• Disadvantages– More expensive than RTT

Expected Transmissions (ETX)

• Estimate number of times a packet has to be retransmitted on each hop

• Each node periodically broadcasts a probe– 802.11 does not retransmit broadcast packets

• Probe carries information about probes received from neighbors

• Node can calculate loss rate on forward (Pf) and reverse (Pr) link to each neighbor

• Select the path with least total ETX

Expected Transmissions (ETX)

• Advantages– Low overhead– Explicitly takes loss rate into account

• Disadvantages– Loss rate of broadcast probe packets is not

the same as loss rate of data packets• Probe packets are smaller than data packets• Broadcast packets are sent at lower data rate

– Does not take data rate or link load into account

Outline

• DSR Revisited




• Conclusion

LQSR

• DSR-like– Route Request, Route Reply and Route Error– Link-quality metrics is appended

• Link-state-like– Link caching (not Route caching)– Reactive maintenance

• On active routes– Proactive maintenance

• Periodically floods RouteRequest-like Link Info

Outline

• DSR Revisited




• Conclusion

Mesh Testbed

Approx. 61 m

Appr

ox.

32 m

23 Laptops running Windows XP. 802.11a cards: mix of Proxim and Netgear.

Diameter: 6-7 hops.

Link bandwidths in the testbed

0

5

10

15

20

25

30

0 5 10 15 20 25 30Higher Bandwidth (Mbps)

Low

er B

andw

dith

(Mbp

s)

• Cards use Autorate •Total node pairs: 23x22/2 = 253

• 90 pairs have non-zero bandwidth in both directions.

Bandwidths vary significantly; lot of asymmetry.

Experiments

1. Bulk-transfer TCP Flows

2. Impact of mobility

Experiment 1

• 3-Minute TCP transfer between each node pair– 23 x 22 = 506 pairs– 1 transfer at a time– Long transfers essential for consistent results

• For each transfer, record: – Throughput– Number of paths

• Path may change during transfer

– Average path length• Weighted by fraction of packets along each path

Median Throughput

0

200

400

600

800

1000

1200

1400

1600

HOP ETX RTT PktPair

Med

ian

Thro

ughp

ut (K

bps)

ETX performs best. RTT performs worst.

Why does ETX perform well?

0

0.2

0.4

0.6

0.8

1

0 2000 4000 6000 8000 10000

Throughput (Kbps)

Cum

ulat

ive

Frac

tion

ETX

HOP

ETX performs better by avoiding low-throughput paths.

Impact on Path Lengths

0

1

2

3

4

5

6

7

8

0 1 2 3 4 5 6 7 8

Path Length with ETX

Path

Len

gth

with

HO

P

Path length is generally higher under ETX.

Why does RTT perform so poorly?

RTT suffers heavily from self-interference

Median Number of Paths

0

5

10

15

20

25

HOP ETX RTT PktPair

Num

ber

of P

aths

Why PktPair gets worse?

PktPair

0

2000

4000

6000

8000

10000

12000

0 1 2 3 4 5 6 7 8

Average Pathlength (Hops)

Thro

ughp

ut (K

bps)

ETX

0

20004000

6000

800010000

12000

0 1 2 3 4 5 6 7 8

Average Path Length (Hops)

Thro

ughp

ut (K

bps)

RTT

02000

40006000

800010000

12000

0 1 2 3 4 5 6 7 8

Average Path Length (Hops)

Thro

ughp

ut (K

bps)

PktPair suffers from self-interference only on multi-hop paths.

Summary of Experiment 1

• ETX performs well despite ignoring link bandwidth

• Self-interference is the main reason behind poor performance of RTT and PktPair.

Experiment 2

Approx. 61 m

Appr

ox.

32 m

Median

0

100

200

300

400

500

600

HOP ETX

Metric

Med

ian

TCP

Thr

ough

put (

Kbp

s)

Outline

• DSR Revisited




• Conclusion

Conclusions

• ETX metric performs best in static scenarios

• RTT performs worst

• PacketPair suffers from self-interference on multi-hop paths

• Shortest path routing seems to perform best in mobile scenarios

Documents

Comparison of Routing Metrics for Static Multi-Hop Wireless Networks