An Incentive Mechanism Based On Reputation In P2P Networks

Recently, P2P networks play an important role in network. However, there exist a large amount of malicious actions in P2P networks resulting from the lack of reputation system. For example, peers are willing to share resources each other and may withdraw from networks at any time. Thus, the system is difficult to guarantee effectiveness and availability. On the other hand, the majority of peers have not activity to contribute resource or provide services to communities due to anonymous participation and voluntary resource contributions in P2P system. Then, this leads to the problem of free-riding. As a result, all of the existing models do not almost take different factors into account when building the trust values. Especially, the trust models in the existing system can not reflect dishonest feedbacks, punishment of the malicious actions, and the incentive mechanism. In P2P networks, free-riding behavior affects seriously the available resources of P2P networks and degrades the overall capability of the services in P2P networks. As a result, solving the free-riding phenomenon is a hot and difficult problem.This thesis presents an incentive mechanism based on reputation values. The mechanism calculates different values between direct and indirect reputation to encourage nodes to provide services for the system in order to maximize their interests. So we can perform the actions to encourage cooperation nodes and punish uncooperative nodes. It makes the phenomenon of "Hot Spots" be effectively alleviated by using load balancing. Furthermore, it can be used to Bloom Filter mechanism to identify and report the malicious nodes rapidly, using the different services to make the uncooperative nodes be excluded from the edge of the system and use BNS mechanism to ease the pressure of backbone network.The analysis and simulation test results show that the mechanism presented can improve overall system availability by identifying the free-riding nodes and malicious nodes in effect and promoting the cooperation between nodes. Meanwhile, the mechanism has good robustness in the aspects of the free-riding behavior inhibition, the alleviation of the "Hot Spots" problem and the identification of the malicious nodes.