On Data-Centric Trust Establishment in Ephemeral Ad Hoc Networks
Maxim Raya, Panos Papadimitratos, Virgil D. Gligor, Jean-Pierre Hubaux
INFOCOM 2008
Ephemeral networks
Definition• No prior associations• Short-lived contacts• Volatile environment• Example: VANET
Trust Properties• Trust in entities can be pre-
established• Reputation is hard to build• Trust in data is important
Trust establishment needs to be rethought
2
What is data trust?
Data Trust in Networks
• Packet forwarding• Security associations• Reputation
AM
B
Data disseminationInsufficientHard
4
Traditional ad hoc networks Ephemeral networks
Data Trust = Entity Trust Data Trust = F(Entity Trust, context)
Event-specific trust
Dynamic trust metricSecurity status
)),(( jkvf ),( jkl v )( kvs
)),(),),((),(( jkljkk vvfvsF
AC
B
M
General FrameworkTrust Computation
Weights (data-centric trust levels)
( )kv is the default trustworthiness
LocationTime
Event reports of type
from nodes jkv
jke
AC
B
M
General FrameworkEvidence Evaluation
( )jBF e
Decision Logic
Evidence Evaluation Output: Decision on Reported Event
Evidence
Event reportsof type
from nodes jkv
jke
( )jCF e ( )jMF e
Decision Logics (1)
• Most trusted report
Decision Logics (2)
• Most trusted report• Weighted voting
Decision Logics (3)
• Most trusted report• Weighted voting• Bayesian inference– Takes into account prior knowledge
Decision Logics (4)
• Most trusted report• Weighted voting• Bayesian inference• Dempster-Shafer Theory– probability is bounded by belief and plausibility– Uncertainty (lack of evidence) does not refute nor
support evidence
Decision Logics (4)
• Most trusted report• Weighted voting• Bayesian inference• Dempster-Shafer Theory
1( ) ( )
K
i k ik
bel m
11
basic belief assignment
( ) ( )ik i km F e trust level
eventreport on event
Decision Logics (4)
• Most trusted report• Weighted voting• Bayesian inference• Dempster-Shafer Theory
1 2, :
1 21 2
, :
( ) ( )
( ) ( )1 ( ) ( )
q r i
q r
q rq r
i iq r
q r
m m
m mm m
12
Dempster’s rule for combination:
1( ) ( )
K
i k ik
bel m
supporting evidence
conflicts
Case Study: VANET
13
Data Trust
Decision on event
Performance comparison
• MATLAB and ns2• 100 simulation runs• 95% confidence intervals• Broadcast environment
Effect of Data Trust (1)
Honest nodes (0.8) are more trustworthy than attackers (0.6)
Effect of Data Trust (2)
Honest nodes (0.6) are less trustworthy than attackers (0.8)
Effect of Uncertainty
Honest nodes (0.4) are more trustworthy than attackers (0.2)
Evolution in Time
Highway scenario; 50% of reports are false (received first);Honest nodes (0.8) are more trustworthy than attackers (0.6)
Conclusions
• Trust in traditional MANETs pertains exclusively to entities
• Data-centric trust is more representative and useful in ephemeral networks (e.g., VANETs)
• Several decision logics can be used, notably:– Bayesian inference if there is prior knowledge– Dempster-Shafer Theory if there is uncertainty
Effect of Prior Knowledge
10 nodes vs. 50 before
Top Related