Trust and Reputation Algorithms for Unstructured P2P Networks

Reputation-based trust management is an effective method to improve network security in P2P networks. In this paper, we present a trust evaluation model based on parameter-estimation algorithms to enhance the accuracy of trust evaluation in unstructured P2P networks. Trust evaluation includes direct and second-hand trust information. It estimates and updates the trust value and uncertainty of each peer by computing mean value and variance for the collected information. We also take an incentive and punishment mechanism to stimulate peers collaboration. Simulations are performed to show the effectiveness of our model. Our model can restrain more malicious behaviors than EigenTrust.