基于网络全局与搜索局部特性的P2P搜索算法  

Search method taking full consideration of both global and partial features in unstructured P2P networks

作　　者：周慧[1] 杨杰[1] 

机构地区：[1]武汉理工大学信息工程学院,武汉430063

出　　处：《计算机应用研究》2010年第8期3115-3119,共5页Application Research of Computers

摘　　要：现有启发式搜索只考虑了搜索过程中的局部性原理,少数的考虑了网络的全局特征,却没有一个算法能同时利用局部性原理和全局性原理来指导搜索过程。为此提出了一种基于动态拓扑调整的搜索算法GLMW,在考虑网络的幂律特征与异构性的基础上利用搜索过程中存在的局部特征来构建搜索协议。通过在仿真软件上的实验,不同的网络规模下比较分析结果表明,算法可以得到更好的查询效果。在同等网络环境下,GLMW查询到的目标副本文件数较多,延时较短,对网络中高能力节点的利用率高。Most existing heuristic search methods only considered partial features existed in search process,or only paid attentions to the global features in networks. This paper put forward a search method called GLMW. The GLMW retained the advantage of dynamic topology adaptation in conforming to the global features in networks,and added consideration of the partial features in object searching process,optimizing current and subsequent requests according to historical search records and making better use of high capacity nodes to serve search process. The simulation results and analysis show that the GLMW can find more object replicas with lower delays,achieve better utilization of high capacity nodes.

关 键 词：对等网络 全局特征 局部特征 缓存定位 遍历器 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]