P2P环境下的全局信任值分级计算方法  

Hierarchical computation approach to global trust values for P2P environments

作　　者：欧阳竟成 林亚平[1] 周四望[2] 张建明[1] 易叶青[1] 

机构地区：[1]湖南大学计算机与通信学院 [2]湖南大学软件学院,湖南长沙410082

出　　处：《通信学报》2008年第8期38-45,共8页Journal on Communications

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)~~

摘　　要：提出了一种P2P环境下不依赖于任何高信任节点集的全局信任值分级计算模型,并给出了它的分布式实现。该模型引入减法聚类方法将大规模网络从逻辑上划分成若干小网络,并将它们重新组织成为上下两层结构的覆盖网,迭代算法在小规模网络中并行进行。迭代收敛后,按层次合成迭代结果便得到每个节点的全局信任值。理论分析与仿真实验表明,与现有模型相比,该模型在降低计算量与通信开销,加速迭代收敛方面有较大改进。A hierarchical computation global trust model, which did not depend on any set of high trust nodes, was proposed for P2P environments. A distributed implementation of the model was presented. By subtractive clustering, a large network was logically divided into some small networks which were reorganized into an overlay network with two layers in the model. Iterations were concurrently done in these small networks. Global trust values of nodes were gained by aggregating these iteration results according to the order of layers after iterations had converged. Theoretic analyses and simulation results show that, comparing to current global trust models, the proposed model can reduce amount of calculations and overheads of communication as well as converges more quickly.

关 键 词：对等网络 信任 分级计算 分布式散列表 开销 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]