摘要： 针对P2P网络提出一种信任模型。通过区分直接信任和间接信任，使推荐信任度主要依赖节点以往的成功推荐次数而非直接信任度，以抵御拥有较高直接信任度的节点对正常节点进行诋毁或欺骗。为避免非恶意节点因为网络延时等原因导致服务失败而被孤立出网络，引入激励机制和信任重建机制，定期提高非恶意节点的信任值，使其能重新加入网络，从而加强网络的容错性。实验结果表明，该模型能有效保护非恶意节点，孤立恶意节点，使网络具有更好的健壮性。

关键词: 信任模型, 激励机制, P2P网络, 信任重建, 直接信任, 间接信任

Abstract: This paper proposes a trust model for P2P network. By distinguishing direct trust and indirect trust, recommend trust value is got according to the past successful recommendation number but not direct trust, which resists behaviors that some Peers with high direct trust cheat others. In order to avoid non-malicious nodes to be isolated out of the network because of failure to offer services, incentive and trust construction mechanisms are introduced, which improves trust value of non-malicious nodes regularly, so that those nodes can rejoin the network and fault-tolerance of the network is strengthened. Experimental result shows that the model can better protect non-malicious nodes and isolate bad nodes, so that the network has better robustness.

Key words: trust model, encouragement mechanism, P2P network, reconstruction of trust, direct trust, indirect trust

中图分类号: 

• TP393