一种基于最小距离索引的对等网系统搜索方法.docVIP

基于small-world网络的非结构化DHT算法* * 基金项目: 国家自然科学基金项目 周晋+，李衍达 (清华大学自动化系 北京 100084) (zhoujin00@) 摘要：针对非结构化系统搜索效率过低的问题，提出一种基于聚类分布的DHT算法，基本做法是将一个环状覆盖网络划分为两层，下层负责数据管理，上层负责节点路由，每层在局部节点上以聚类方式组织信息。该算法初步解决了已有非结构化算法的数据组织无序性和搜索盲目性。为了进一步减小搜索路径长度，改进算法利用small-world的研究成果，在HUB层中加入指向远距离节点的远程连接，使得严格聚类转为带有随机性的概率化聚类，从而将搜索路径长度控制在数量级。 关键词：Peer-to-Peer；搜索；聚类分布；small-world；DHT 中图法分类号：TP393 A Peer-to-peer DHT Algorithm Based on Small-world Network ZHOU Jin, LI Yan-Da (Department of Automation, Tsinghua University Abstract To address the inefficiency of existing unstructured algorithms, the paper presents a DHT algorithm based on clustering distribution of indices. The algorithm divides a ring-like key space into two layers. The lower layer is responsible for indices management, and upper layer is responsible for routing nodes. On both layers, indices are clustered within each peer. The algorithm solves the problems of present unstructured algorithms in a certain degree, in which the datum is organized in a confused situation and routing is proceed in a disordered manner. Further, enhanced algorithm uses outcome of small-world to cut down latency of routing hops. With enhanced algorithm, a few connections to remote peers are introduced into routing table with a suitable probability. In another word, the criterion of clustering is not of determination, but of probability. Thus, it is possible to route overlay with average latency of hops. Keywords Peer-to-Peer, route, clustering distribution, small-world, DHT 1 简介 近来，Peer-to-peer系统（简称P2P系统）在文件共享和信息搜索等方面得到了越来越多的应用，Morpheus ADDIN REFMGR.CITE RefmanCiteYear2003/YearRecNum14/RecNumIDTextMorpheus/IDTextMDL Ref_Type=Internet CommunicationRef_TypeInternet Communication/Ref_TypeRef_ID14/Ref_IDTitle_PrimaryMorpheus/Title_PrimaryDate_Primary2003/Date_PrimaryReprintNot in File/ReprintWeb_URLu/u/Web_URLZZ_WorkformID35/ZZ_WorkformID/MDL/Cite/Refman[1]的系统报告指出：截止2001年10月26日，用户数量超过470,000位，共享文件总量约360TeraBytes。P2P系统是由一组地位相等的节点构成，节点间可以直接通讯，无需第三方参与。与C/S结