A New Trust Model Based On Reputation And Recommend Evaluation For P2P Networks

Peer-to-Peer network has become one of the most important Internet application systems in the world. However, P2P have the characteristics of openness, anonymity and loosely coupling relationship between nodes, which leads to the lack of an effective mechanism in the network to improve the availability for the overall system. System security is facing enormous challenges since these characteristics will cause unreliable service quality problems:(1) Free-Riding issues (2) Tragedy of Common issues (3) Massive fraud, forgery and other malicious acts.The issue of trust in the P2P network system can be resolved by establishment of a reliable P2P trust model. According to studying and learning from the typical solutions of trust mechanism in China and other countries, in this paper, we will put forward a set of solutions involved P2P trust computing model and P2P trust management model, based on the improvement of typical trust algorithm under distributed environment. Moreover, we have performed simulation test and analysis on the performance and effect of the model.Using social network theory, the model in this paper is able to apply several parameters to simulated network transactions and estimate possible future transactions by calculating the comprehensive trust between nodes. It can dynamically balance between direct transaction experiences and recommendations fedback by friend node and gather comprehensive reputation information to calculate comprehensive trust. Besides, it can avoid the slander and overstating to a recomend node by means of designing recommendation evaluation criteria based on normal distribution, thus presents a new trust model based on reputation and recommend evaluation for P2P networks. Furthermore, it can compute comprehensive trust value between nodes veritably and effectively. As a result, the incentives and punishment mechanism introduced by recommended evaluation criteria is able to deal with the problem of inert nodes and prevent the behavior of malicious node.Following the Terrace mechanism, we designed not only a distribution and storage strategy for trust information of RRE-Trust model, but also a distributed algorithm. We improved the file node's location where the trust information of Terrace topological node stored so as to ensure that the file node of highly-trusted nodes are also stored on the other highly-trusted nodes and can load more in the actual trust management and P2P applications, thereby reaching balance of load. Meanwhile, it ensures the secure storage and access of trust data and enhances the anonymity and security in the transaction of P2P network.