A Trust-Enhanced Topology Adaptation Protocol for Unstructured P2P Overlays

In this paper, a novel probabilistic computational reputation based trust model is firstly introduced, which has the property of separating reputations between providing services and giving feedbacks. And then a topology adaptation protocol for unstructured P2P overlays is proposed based on this trust model. The basic principle of this adaptation protocol is that peers with higher service and feedback trust values have higher probabilities of being accepted as neighbors in the P2P overlay. We show via simulations that the proposed trust model has better accuracy of reputations generated. And the topology adaptation protocol can enhance trusts even further, resulting in fewer malicious transactions when highly dynamic peer behavior patterns exist in the P2P network.