Sustaining Cooperation in
Multi-Hop Wireless Networks
Ratul Mahajan, Maya Rodrig, David Wetherall and John
Zahorjan
University of Washington
Presented by:
Prasad
(Slides in courtesy of: Bin Ni, University of South Carolina)
1
Key words

Wireless network
- characteristic of broadcasting

Multi-Hop
- a packet may be forwarded by multi nodes

Free-rider and Cooperation
- less contribution to the group
- consume more than their fair share of a resource
2
Topics





Problem Definition
The power of Anonymity
The Catch Protocol
Experimental Evaluation
Simulation and Analysis
3
Problem Definition

B can avoid these forwarding loads in
two distinct ways:
- forwarding level : drop packets for forwarding from A
- routing level : refuse to send routing messages that
acknowledge connectivity with A
4
Assumptions



Most nodes are cooperative in that they
run the protocol defined;
Omni-directional radio transmitters and
antennas;
Nodes have unforgeable ID.
5
Power of Anonymity


The goal is to use cooperative nodes to
monitor for the presence of free-riders
and to isolate them.
Two problems
- distinguish them
- signal all of the free-rider’s neighbor
6
Anonymous Challenges and Watchdogs

A watchdog is used;

Anonymous challenge message (ACM)

sub-protocol is defined;
Gateway regularly but unpredictably
sends an anonymous challenge to the
testee to rebroadcast.
7
Anonymous Neighbor Verification



Once free-rider is detected, other
testers must also be informed;
Challenge: the only path must be via
the testee;
Anonymous neighbor verification (ANV)
sub-protocol is defined;
8
Anonymous Neighbor Verification

ANV open
- all the testers become aware of each other via the testee;
- each tester sends a cryptographic hash of a randomly
generated token;
- testee rebroadcasts, other testers take note;

ANV close
- each tester releases its token to the testee if it behaved well;
- the testee broadcasts the token;
- testers compare the received token;
9
Catch protocol

Epoch-Start
- testee broadcasts packet that includes its ID and epoch ID;
- nodes that receives the packets participate as testers for this epoch;

Packet Forwarding and Accounting
- testers run a watchdog to count the number that are correctly
relayed;
- testers run the ACM protocol to estimate true connectivity;

Anonymous Neighbor Verification Open
- tester sends an anonymous packet containing a nonce and a hashed
token to the testee for rebroadcast
10
Catch protocol

Tester Information Exchange
- tester compares the fraction of data packets and challenges;
- one-bit sign:0 for challenges and 1 for data packets;
- send the sign bit and ID to testee for rebroadcasting ;

Epoch Evaluation and ANV close
- each testers determine whether the testee behaviors correctly
- this is done with a pair of statistical test
- both tests pass, the tester releases its token;

Isolation Decision
11
Protocol flow: packets exchange between a tester and a
cooperative (left side) or free-riding (right side) testee.
12
Evaluation





15 PCs equipped with
802.11b
Operating in the ad-hoc
mode
Diameter is between 3 and
5 hops
Length of one epoch is set
to one minute
There are 15 anonymous
ACM messages per epoch
13
Evaluation



Three nodes: the second
one acts as free-rider
The number of epochs
required to detect freeriders in the testbed
versus the fraction of
packets a free-rider
dropped.
Each point is the average
of 10 experiments.
14
Evaluation



Three nodes are freeriders that drop all the
packets ;
Nodes select only
cooperative nodes as
file servers;
It shows that the
cooperative nodes
successfully shut up the
free-riders;
15
Evaluation



We ran two five hour experiments in
which all nodes were cooperative.
Each node repeatedly downloaded les
from randomly chosen servers.
We observed no false positives in the
first experiment and a single false
positive in the second.
16
Evaluation


cheater that uses signal
strength to differentiate
among its neighbors.
For example, when
20% (3) of the nodes
cheat, that probability
is lowered from about
60% to about 10%
when using the highest
quality links.
17
Simulation and Analysis


the impact of high wireless loss rates on
Catch is quite small.
Catch seems to be robust, working well
in both high and low density networks.
18
Future work


Collusion of the free riders
Taking specific steps in Catch to
discourage signal strength based cheats
19
Thanks!
Any question?
20