Research On Trust Model Of P2P Based On Peer's Reputation Level

With the deepening of P2P technology, P2P as a distributed network technology can be widely used in many fields, especially in resource sharing. However, due to the existences of network security problems which are inherent in network have becoming the key factor of restricting the P2P resource sharing. The research of trust model provides a program that can effectively solve the security of P2P resource sharing. The main function of the trust model is to complete the assess of nodes in the P2P network, and provides the trusted relationship for the safe interaction of node-users and promotes the node-users in the network participate in the net actively.. But the present trust model doesn't effectively restrict the free-riding behavior of the node in the network. In response to the above problems, this paper proposed a new trust model—Base on multilevel reputation system of P2P network security trust model. This paper focus on the following aspects:①The analysis of the current P2P trust model. On the base of studying and handling the feature of the P2P network technology and the network topology, we analyzed the trust feature and the structure of trust model, classified the existing trust model, and summarized the main current model's strengths and weaknesses, which provides the theory basis for constructing the multilevel P2P network security trust model in this thesis.②In connection with the fact that the quality of service in existing network models didn't have the uniform standards, we propose a new method that use the setpairanalysis (SPA) method to classify the quality of service. And finally, we use the weighted calculation to get the trust value.③In consideration of the flooding of nodes free-riding act, we propose a new trust model that based on multilevel reputation (BMLR). Our trust model proposes an incentive mechanism that can be used to award the node which contributes to the network.④The simulating experiment. Through the simulating experiment, we verified the BMLR trust model which can effectively control the free-riding node in the network. Finally, base on the summary of this paper, we propose numbers of problems that need further study and discussion.