The Research And Implement Of A Mechanism For Trust Management Based On Fuzzy Theory In P2P Networks

P2P(Peer-to-Peer) is presently one of the most important hotspots in the network, and what strongly affects the traditional network is that it adopts a distributing network structure, and focuses on its content. Each peer in the network is both client and server. It has special charms in the application areas like file sharing, distributed computing, cooperation system, and deep search engine, etc. But its nature of openness, equality, autonomy, and free of supervision also lead many new security problems such as a lot of malicious peers and Free-rider peers. It is too hard for us to asure the security of P2P networks if these problems can't be well solved. Therefore, it is necessary to construct an effective and secure trust management mechanism and research the key technology.This paper focuses on trust management mechanism. Aiming at the problems of malicious peers and free-rider peers, it proposes TMFT mechanism which is shorted by the title of this paper. The main features of this paper are blow.First, TMFT which effectively combines trust management with incentive mechanism can more conveniently manage P2P networks. In order to solve problems of malicious peers and Free-riding, most existing models use either turst management or incentive mechanism. As we all know, trust management can effectively control the behaviors of resource or service providers, and incentive mechanism can effectively control the behaviors of resource or service consumers. Therefore, this paper combines the two mechanisms, and it can effectively control the behaviors of providers and consumers.Second, TMFT proposes a novel model for trust management based on fuzzy theory in P2P networks, in which the reliability of recommendation trust can be computed by aggregating similarity and bayes estimate. It not only can efficiently deal with subjectivity, fuzziness and uncertainty, but also can prevent the collusion in malicious nodes and spy actions, and finally malicious peers will be out of the networks.Last, TMFT proposes a method based on concurrent signatures and a new arbitration mechanism, which can solve fairness problems in the exchanges. Besides, it introduces two ways of accumulation of virtual capital. So it not only can provide better fairness of exchanges, but also can reach the aim of incentive.Finally, this trust model is implemented by simulation experiments which show that this proposed trust model not only can distinguish the malicious peers effectively, but also increase the good files ratio in P2P download files.