Research On Reputation Model Based On Repeated Game Theory And Punishment Mechanism In P2P Environment

In recent years, P2P network has been developing rapidly in Internet, owing to its advantage in sharing resources, including information, files, bandwidth, storage spaces, processor cycles, and so on. P2P network is a new technology, which has a broad development space. Each node both a client and a server in the network is equal and the network is not managed by any special authoritative node. The characteristics of distribution and self-organization cause the wide application of P2P network, however, which also makes the network security issues prominent.The behavior that each rational node maximizes the resources of other nodes and makes least contributions to the network determined by its selfishness is becoming more serious. Moreover, a large number of unreliable resources and servers exist in network, and the node has abundant non-cooperative behaviors and frauds too. Thus, it is necessary to establish an effective mechanism to constrain nodes selfishness and encourage collaboration among nodes. For this purpose, a trust model is established in this paper.To deal with the shortcomings in high time complexity and network overhead in calculating nodes' trust degree, as well as to take into account the impact of rational nodes ratio and nodes patience degree on the trust model, the punishment mechanism is utilized to establish the P2P trust model based on the repeated game theory. The trust mechanism is described and the repeated trade actions are analyzed by using repeated game theory. The collaborative equilibrium theorem considering the rational nodes ratio and nodes patience degree is proposed and proved, from which a collaboration condition is concluded and the nodes should adopt the collaboration strategy for fear of being punished under this condition. The experimental results show the effectiveness of this trust model in promoting the nodes collaboration.