LDTS: A Lightweight and Dependable Trust System for Clustered Wireless Sensor Networks

1. LDTS: A Lightweight and Dependable Trust System for Clustered Wireless Sensor Networks Authors: Xiaoyong Li, Feng Zhou, and Junping Du Presented by: Ting Hua

2. Outline • Motivation • Clustered WSN Model • Lightweight Scheme for Trust Decision-Making • Theoretical analysis and evaluation • Simulation-based analysis and evaluation • Conclusion

3. Motivation • Limited work focus on • Resource efficiency of clustered WSNs • fail to consider the problem of resource constraints of nodes • used complex algorithms to calculate nodes’ trustworthiness • Dependability of the trust system itself • Current: collect remote feedback and then aggregate s such feedback to yield the global reputation for the nodes • Problem: How about open or hostile WSN environment contains a large number of undependable (or malicious) nodes?

4. Outline • Motivation • ClusteredWSNModel • Lightweight Scheme for Trust Decision-Making • Theoretical analysis and evaluation • Simulation-based analysis and evaluation • Conclusion

5. Clustered WSN Model • Nodes • CH: cluster head • CM: cluster member • BS: base station • Communications • Inter-cluster: A CM can communicate with their CH directly. • Intra-cluster: A CH can forward the aggregated data to the central BS through other CHs.

6. Outline • Motivation • Clustered WSN Model • Lightweight Scheme for Trust Decision-Making • Theoretical analysis and evaluation • Simulation-based analysis and evaluation • Conclusion

7. Trust Decision-Making at CM Level • Decision making: past interaction records? • Yes: CM-to-CM Direct trust degree (DTD) • # of successful and unsuccessful interactions • Interaction: cooperation of two CMs, e.g., node x sends a message to CH ivia node y • Successful: node y forwarded such message to CH • Unsuccessful: • No retransmission of the packet within a threshold time • Overheard packet is illegally fabricated • No: CH-to-CM Indirect trust degree (ITD) • send a feedback request to CH

8. CM-to-CM Direct Trust Calculation a window of time # of successful interactions of node x with y # of unsuccessful interactions of node x with y strict punishment for unsuccessful interactions

9. CH-to-CM Feedback Trust Calculation # of positive feedback # of negative feedback Assumption: CH is trustworthy within its cluster!

10. Trust Decision-Making at CH Level • Decision making: calculate for direct trust and feedback trust simultaneously • CH-to-CH direct trust • # of successful and unsuccessful interactions • BS-to-CH feedback trust • BS periodically asks all CHs for their trust ratings on their neighbors. • CH send a feedback request to BS

11. CH-to-CH Direct Trust Calculation a window of time # of successful interactions of CH i with CH j # of unsuccessful interactions of CH i with CH j strict punishment for unsuccessful interactions

12. BS-to-CH Feedback Trust Calculation feedback of CH k toward CH j quality of feedback # of positive feedback # of negative feedback

13. Self-Adaptive Global Trust Aggregation at CHs BS-to-CH feedback trust CH-to-CH Direct Trust # of positive feedbacks # of successful interactions increasing α, Φ(x) quickly approaches 1

14. Outline • Motivation • Clustered WSN Model • Lightweight Scheme for Trust Decision-Making • Theoretical analysis and evaluation • Simulation-based analysis and evaluation • Conclusion

15. Dependability Analysis Against Malicious Attacks

16. Dependability Analysis Against Malicious Attacks

17. Dependability Analysis Against Malicious Attacks

18. Dependability Analysis Against Malicious Attacks

19. Dependability Analysis Against Malicious Attacks

20. Dependability Analysis Against Malicious Attacks

21. Dependability Analysis Against Malicious Attacks

22. Dependability Analysis Against Malicious Attacks

23. Dependability Analysis Against Malicious Attacks

24. Dependability Analysis Against Malicious Attacks

25. Communication Overhead Analysis and Comparison Assume: Network consists of m clusters (including the BS) average size of clusters is n (including the CH of the cluster) communication overhead of one node # of CM send n requests and receive n responses

26. Storage Overhead Analysis and Comparison

27. Outline • Motivation • Clustered WSN Model • Lightweight Scheme for Trust Decision-Making • Theoretical analysis and evaluation • Simulation-based analysis and evaluation • Conclusion

28. LDTS Simulator and Environment

29. Overhead Evaluation and Comparison

30. Overhead Evaluation and Comparison

31. Dependability Evaluation and Comparison

32. Dependability Evaluation and Comparison

33. Outline • Motivation • Clustered WSN Model • Lightweight Scheme for Trust Decision-Making • Theoretical analysis and evaluation • Simulation-based analysis and evaluation • Conclusion

34. Conclusion • Lightweight trust evaluating scheme • cooperations between CMs • cooperations between CHs • Dependability-enhanced trust evaluating approach • cooperations between CHs • Self-adaptive weighting method • CH’s trust aggregation