Abstract: P2P network has the characteristics of dynamic, self-organization and anonymity, which leads to the existence of malicious behavior of some nodes in the network. In order to ensure that the node can provide reliable resources and good service, this paper proposes a P2P network trust model based on time series, building trust relationship between peers. This trust model introduces a time decay function when calculating the direct trust and indirect trust, and the recent trust value is small, which has great influence on the overall trust degree. The trust value of cluster head nodes and cluster nodes are updated intime using feedback control mechanism, so that the value of the overall trust has timeliness, the evaluation of node trust is more accurate and more reliable. The experiment proves that this trust model can provide safeguard and improve credibility to P2P Network because it has better dynamic self-adaptability, and better ability to check the malicious peers than DyTrust.

Key words: P2P network, comprehensive trust model, time series, human trust relationship, direct trust, indirect trust

摘要： P2P网络具有动态性，自组织和匿名性等特点，从而导致网络中某些节点存在恶意行为，为了保证节点能提供可靠资源和良好服务，提出一种基于时间序列的P2P综合信任模型，使对等节点之间建立信任关系。该模型在计算直接信任度和间接信任度时引入时间衰减函数，近期的信任值衰减小，对整体信任度的影响大。通过反馈控制机制动态更新簇内节点和簇头节点的信任度，使整体信任值具有时效性，对节点信任度的评价更精确和更可靠。实验证明，该信任模型较DyTrust模型降低了节点之间的交互时延，稳定性和动态自适应性更好，检测恶意节点的能力更强，提高P2P网络的可信度，对P2P网络的安全提供有力保障。

关键词: P2P网络, 综合信任模型, 时间序列, 人类信任关系, 直接信任, 间接信任