GARM: A group - Anonymity reputation model in peer-to-peer system

When Peer-to-Peer reputation systems are optimal for attracting peers to contribute resources and combating the rise in "Free-Riding", they still cannot avoid facing some malicious attacks such as "Slander", "Collusion", and "Sybil Attack". In this paper, reputation mechanism and anonymity in Peer-to-Peer systems are both considered. A group-anonymity reputation model without CA is proposed, for the purpose of impelling the P2P system to the maximization of resource utility. By way of contrasting to other reputation systems, peers in this model do not have a global reputation but a group-anonymity reputation instead. In addition, peers are divided into distinguishing groups, among which anonymity is achieved - by preventing querying peers from getting the information of the providing peers. And after each interaction between the service peer (namely, the resource provider) and the consumer one, the latter will submit its evaluation about the provider to the system. As the anonymity of the resource provider is protected, some malicious behaviors could not exercise a great influence on this group-anonymity reputation model.