Dual-EigenRep: A Reputation-Based Trust Model for P2P File-Sharing Networks

Trust research has become a key issue in the last few years as a novel and valid solution to ensure the security and application in peer-to-peer (P2P) file-sharing networks. The fundamental challenges lie in what transaction behaviors are most suitable for the measure of reputation and how to resist various threats of manipulation by strategic participants. In this paper, we propose a new trust model based on the technology to calculate the eigenvectors of trust matrices. We define recommended reputation value and recommending reputation value to evaluate the historical transaction behavior for each peer and make them rely on each other tightly. After certain transactions our algorithm makes the entire network from different trust communities, which are corresponding to normal and different types of malicious peers. The reputation values in normal peer community become higher by performing normalization process, thus the behavior of malicious peers will be confined largely. Experiment results confirm the efficiency of our model against the threats of exaggeration, collusion, disguise and single-behavior.