원문정보
초록
영어
Trust system plays a more and more important role in giving nodes incentives to cooperate in packet forwarding especially for wireless ad hoc networks. However, most existing works in this field either lack rigorous analysis of cost of their methods or have analysis in unrealistic models, which will clearly damage their effectiveness in real applications. In a previous work, Theodorakopoulos and Baras [1] develop a novel formulation of trust computation based on ordered semirings. In this paper, based on their work we proposed FROST (FRiendship and Ordered Semirings based Trust system) for wireless ad hoc networks. FROST introduce notion of friendship to reduce the trust table size and overhead while building and maintaining the trust system in large scale networks. Moreover, FROST use a more effective decaying model to enable the trust system to be more adaptive to the changing environment. Rigorous analysis and in-depth simulation show FROST has a good performance and can quantitatively measure reputation and defend trust system against malicious attacks.
목차
1. Introduction
2. Related Works
3. Design of Novel Trust System
3.1. Reputation computation engines
3.2. Decaying model
3.3. Friendship model
4. Process of Merging Trust Semiring, Decaying Model and FriendshipModel
5. Attacks and Analyses
5.1. Black/Gray hole attack
5.2. Conflicting Behavior Attack
5.3. Selective misbehavior
5.4. Bad Mouthing Attack
5.5. On-off Attack
6. Performance Analysis
6.1. Performance comparison between the new approach and reference methods
6.2. Trust value computations after D is introduced
7. Conclusions
Acknowledgements
References