The Research Of A Reputation Model Based On Hybrid Hierarchical P2P Network

In recent years, the security problem in P2P networks has taken more and more attention of the industry and academia. The main solution to solve this problem is to establish the trust mechanism. At present, how to identify the malicious peers and inhibit their misbehaviors has become the focus of the research in the trust mechanism domain.Referring to the trust relationship of the communication network of human society and the characters of the hybrid hierarchical P2P network, a new reputation model named ClusterRep based on hybrid hierarchical P2P network is proposed in this thesis. In this model, the nodes are divided into groups according to their interests. The trust relationships are classified into three tiers:the trust relationship between peers, between groups, and between peers and groups. The trust values are evaluated via combinations of two types of values:the direct and the indirect trust value. The former one is taken from the first trust relationship, and the latter one is from the other two trust relationships. After the modeling and simulation of ClusterRep, we conclude that ClusterRep can effectively identify malicious peers, and enhance the ratio of the success query and the satisfaction level.