P2P Trust Model Based On Gray Theory

With the development of Internet, peer to peer network (P2P network for short) becomes one of the hot research areas at present. Taking full advantage of users'resources in the network, P2P technology improves the accuracy of search and location of resources as well as the downloading speed, so it is widely used in many areas. In addition to its vast potential for future development, P2P network faces great challenges, so the network security is of great importance. Therefore, while providing convenience for the users, how to build a reasonable trust management system in P2P network becomes very important, so that nodes in the network can offer more effective services.What trust reflects is the comprehensive evaluation of a node to another node on its behavior and ability. In addition, the trust between the nodes in P2P network is the premise to their mutual cooperation, mutual encouragement, and resources sharing, and it has a direct impact on the overall performance of P2P systems. The issue of trust can be resolved by building a reliable trust management model. Consequently, this thesis mainly focuses on the study of P2P network trust model, refers to the trust mechanism of human society, and proposes P2P trust model based on gray theory.The thesis mainly consists of the following sections:First, it gives a detailed description on the basic knowledge of P2P network, and inquires into some typical trust models and corresponding research methods.Second, as for the problems such as the fuzzy, inaccurate trust values in the network and great cost on network operation, domain of interest and calculation method of trust values based on gray theory can be applied. However, it is their resources that interest the user nodes, each of which itself also has its own interests and hobbies. Therefore, the nodes with the same interests divided into one domain, the search cost will be effectively reduced, and success potential improved, too. To calculate the comprehensive trust values, we combine the direct trust values and the indirect ones. The direct trust values can be calculated through the sum of weighted evaluation factors, and the indirect ones through gray correlation. Meanwhile, evaluation factors introduced can describe the feature that trust decays with passing time. In this way, the comprehensive trust values achieved will be more accurate, which provides successful transaction with reliable security. Third, the model conducts a security analysis on the attack of the malicious nodes to the trust model, from the aspects such as recommended cheating behavior and collaborative fraud. In addition, the encouragement of trust model will be accomplished through the interest factors in the local trust relational table; the thesis also analyzes the "free rider" behavior, and puts forward a counter-proposed plan.Fourth, simulation experiment on the model proposed in the thesis is carried out, and the results obtained are analyzed.