Skip to Main Content
Reputation-based trust management mechanism has been identified as a viable solution to the security problems of peer-to-peer (P2P) networks in the last few years. The authors' goal is to find a better model for good peers to withstand malicious activity of single malevolent peers and their collusions as well. In this article, the authors propose a novel trust model, based on parameter-estimation techniques named P-Trust to cope with misbehaviour in unstructured P2P networks. In this model each peer maintains direct trust information of others. The reputation system makes one peer evaluate other peers by the self-experiential and second-hand information. Meanwhile, incentive and punishment mechanism is taken to motivate peers to provide true information. The authors test and analyse the performance of a P-Trust against misbehaviour and the implementation costs. As a result, the authors' model shows more effectiveness and robustness than the previous model.