Reputation computation based on new metric in P2P network

For most of P2P file sharing trusted download, the current researches of the evaluation of the trusted peers based on reputation simply rely on the history of interaction and are not paid enough attention on the time factor. The other behaviors like recommendation are also ignored. The attack actions in reputation evaluation were discussed, and the recommender behavior in P2P network was considered separately, which made up for computation based on the interaction behavior in computation of reputation. Secondly, for the positive and negative evaluation, different time-decline coefficient was added in the method to distinguish the malicious node named smart node which acts as a malicious node at times. Simulation experiments show that this proposed method not only increase the good files ratio in P2P download files, but also restrain the collusion in malicious nodes, especially the spy action.