Research On Trust Management Model In P2P Network

With the rapid development of peer-to-peer technology, it brings us not only opportunities, but also threats. Because of lacking centralized management, all kinds of threats appeared in the P2P networks are becoming the main obstacle that prevent the system from operating normally, which is mainly caused by some nodes'malicious behaviors. How to recognize the malicious nodes, preventing them from putting their abnormal plans in force is becoming a new problem. And nowadays, the research on nodes'trust relationship has attracted more and more researchers'attentions. Reputation is a comprehensive assessment of node's past behaviors.The trust model in this paper is built through simulating social network, the reputation of the peer determines whether the transaction can process or not, a peer's reputation mainly comes from other nodes'weighted trust feedback. Besides, some important aspects which affect peer's reputation are also taken into account in the process of reputation computation, such as download volume, upload volume, evaluation volume, online time, time attenuation coefficient and so on. Due to these parameters, the reputation can reflect the actual situation of peer better, and what's more, it also provides a kind of incentive mechanism which will inspirit all users to join the system actively.The paper also presents a safe data storage scheme. The peer's content value of each transaction is stored on its local database while its reputation and trust in formations are stored in other peers based on DHTs mechanism (distribution hash tables) with a reasonable redundancy. With this method, the peer can not only access data more rapidly and efficiently, but also prevent tampering attacks effectively.