双向主从式Chord资源搜索算法的研究  

Research of the Bidirectional Master-Slave Chord Resource Search Algorithm

作　　者：许智宏[1] 张竹楼[1] 郭浩[1] 李雪原[1] 

机构地区：[1]河北工业大学计算机科学与软件学院,天津300401

出　　处：《计算机工程与科学》2011年第11期80-83,共4页Computer Engineering & Science

摘　　要：Chord是一种结构化的P2P网络模型,它具有速度快、无需中心控制、可扩展性强、负载平衡、高容错性能等优点。但是,Chord查找算法为单向查找,当目的节点与当前节点距离较远时,需经多次跳转,增加了路由延迟;Chord中能力较弱的节点来负责系统中大量的查询和下载,以及节点随时加入或离开系统的频繁变迁情况,这样会造成网络查询效率明显下降。改进的算法即双向主从式Chord算法支持双向搜索,并将网络中的节点分为超级节点和普通节点,由评估结果值较高的超级节点组成Chord主环。通过实验证明,改进算法有效地减少了路由跳数,降低了网络延迟。Chord is a structured P2P network model, it has the advantages of high speed, needing no center control, extensibility, load balance, fault tolerance, etc. But, as Chord is a one-way search algorithm, when the destination node is far from the current node, the routing delay increases because of multihops. And, low ability nodes with responsibility for large numbers of search and download tasks, and the nodes join in or leave the system momentarily and frequently, which can decrease the network efficiency obviously. An improved algorithm called the bidirectional master-slave chord algorithm supports two-way search, and all the network nodes are divided into super nodes and common nodes, and the super nodes with higher evaluation results construct the main ring of Chord. The experimental results show that the improved algorithm can reduce routing hops and network delay effectively.

关 键 词：对等网络 分布式哈希表 CHORD 资源搜索 

分 类 号：TP393.2[自动化与计算机技术—计算机应用技术]