摘要
现有信任模型对节点行为的突然改变不能做出迅速地反应,动态性适应能力支持不足.引入反馈控制机制,提出了一种P2P网络环境下的全局信任模型,并给出模型的分布式实现策略.该模型排除了P2P网络中很难确定的高信任节点集合,并通过对反馈信息的有效聚合增强信任系统的动态适应能力,有效地遏制恶意节点的动态摇摆行为,提高了P2P网络中共享文件的下载成功率.仿真结果显示,在有50%的动态摇摆恶意节点的P2P文件共享网络环境下,真实文件下载的平均成功率是用于比较现有模型的1.5倍.
Existing trust models can not provide adequate reaction to quick changes in peers' behavior, and thus lack dynamic adaptability. A global trust model integrating feedback control mechanism for P2P networks was proposed, which includes a distributed implementation. The model excludes pre-trusted peers, who are hardly determined in P2P systems, and enhances the dynamic adaptability of trust systems by aggregating feedback information. It can efficiently suppress malicious peers' oscillating behavior and improve the successful download ratio of files in P2P file-sharing networks. Simulation results show that the average download ratio of authentic files is 1.5 times bigger than the global trust models under comparison when malicious oscillating peers is 50 percent of the population in P2P file-sharing networks.
出处
《湖南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2008年第8期68-72,共5页
Journal of Hunan University:Natural Sciences
基金
湖南省自然科学基金资助项目(06JJ20049)
湖南省教育厅科研基金资助项目(03C162)
关键词
对等网络
信任
摇摆行为
分布式哈希表
Peer-to-Peer
trust
oscillating behavior
distributed hash table