A New Reputation Management Mechanism Based on Bi-Ratings in Peer-to-Peer Systems

The efficiency of reputation system depends on the quality of feedbacks. However current reputation models in peer-to-peer systems can not process such strategic feedbacks as collusive ratings or no ratings attacks. Also there is unfairness for the blameless peers in these models. Pointing to these problems, a new reputation management mechanism is proposed. In this mechanism there are two metrics used to evaluate peers: feedback and service trust. Service trust shows the service reliability of peers. Feedback trust can reflect credibility of peers when reporting ratings. After a transaction both service consumer and provider submit ratings to report the quality of this transaction. According to these two ratings, service trust of sever and feedback trust of consumer are separately updated, furthermore the former is closely related to the latter. Complementary to the reputation model, a punishment mechanism is proposed to prevent malicious peers from iteratively exerting bad behaviors or not submitting ratings. Although both the partners are punished when two ratings disagree with each other, only the service trust of provider may be decreased but its feedback trust keeps constant. Likewise, feedback trust of reporter will be decreased while its service trust does not change. Simulation shows that the proposed approach can effectively resist aforesaid malicious attacks and mitigate unfairness.