无结构对等网络中基于信誉的激励模型研究

Research on the Incentive Model Based on Reputation in Unstructured Peer-to-peer Network

将社会网中的信任机制引入到无结构对等网络中,通过社会网相关理论在节点间构建信任关系.把网络中每个节点同时拥有的"服务贡献者"和"服务消费者"这两个不同角色,分别采用不同的信誉值进行衡量与评估,并以此构建激励模型.在推荐信誉值的评估中,进行了针对恶意节点虚假推荐值的去噪声处理,提高了信誉值评估的抗干扰性.实验结果表明,该模型能够有效抑制搭便车和恶意节点等不合作节点的行为,在保障系统公平性的同时,提高了系统的可靠性与服务质量.

This paper introduces the trust mechanism of social network into unstructured peer-to-peer （P2P） network, and establishes the trust relationship between related peers with the social network theories. Since each peer in the network has two roles of ＂service contributor＂ and ＂service consumer, two different reputation evaluation metrics are adopted and assessed in the incentive model built in this article. In the process of evaluating the value of recommendation reputation, those fake data from malicious peers are treated exclusively to solve the problem of noises so as to improve the interference immunity on reputation evaluation. Simulation results show that the proposed trust model can restrict effectively the action of free-riders and malicious peers, ensure the fairness in the system, and improve the system＇s reliability and quality of service.