Abstract: Through lack of strict identification and trust mechanisms, there are a lot of malicious peers in current P2P network. This paper introduced several typical trust models and malicious behavior patterns, then analyzed the characters of these trust models and their resistance to difference malicious behaviors by simulations. It shows that trust models based on feedback can effectively protect the network from many kinds of malicious behaviors.

Key words: trust model, malicious behavior, Peer-to-peer

摘要： 在目前已广泛应用的P2P网络中，由于缺乏严格的身份验证和信任机制，存在着许多欺诈等恶意行为，系统的有效性和可用性难以保证。本文总结了P2P网络中的信任模型和恶意行为模式，并针对不同的恶意行为分别对几种典型的P2P信任模型进行模拟试验，对比分析了P2P信任模型的特点和对各种恶意行为的抑制作用。结果表明基于反馈的信任模型能够有效地抵御多种恶意节点的破坏、提高网络服务的有效性。

关键词: 信任模型, 恶意行为, 对等网络