A Decentralized Scheme For Reputation Management In P2p Virtual Communities

Jingtao Li,Xueping Wang,Jianqing Ma,Xuefeng Jia,Qian Wang,Gendu Zhang
2006-01-01
Abstract:In peer-to-peer (P2P) virtual communities, a peer needs to interact with unknown peers for the services provided. Peers have to manage the risk involved with the interactions without prior knowledge about each other's reputation.PerSimTrust, a personalized reputation management scheme, is proposed for P2P virtual communities. Each peer uses a personalized similarity measure to rate the credibility of another peer's recommendation. The trust value of a peer is computed by aggregating similarity-weighted recommendations of the peers who have interacted with the peer. PersimTrust also considers decentralized trust data management schemes to which not enough importance is given before.Theoretical analyses and experiments show that the scheme is robust under a variety of conditions where malicious peers cooperate in an attempt to deliberately subvert the system, and effectively decreases the number of inauthentic files downloaded in a simulated P2P rile-sharing community.
What problem does this paper attempt to address?