具有位置感知和语义特征的P2P网络模型  

Semantic and Location-Aware Based Unstructured P2P Network

作　　者：孙新[1] 刘玉树[1] 刘琼昕[1] 郑军[1] 

机构地区：[1]北京理工大学计算机学院,智能信息技术北京市重点实验室,北京100081

出　　处：《电子学报》2010年第11期2606-2610,共5页Acta Electronica Sinica

基　　金：国防预研项目(No.104080402);北京市重点学科建设项目;北京理工大学科研基地科技支撑计划项目

摘　　要：非结构化P2P网络中常用的泛洪搜索方法造成了严重的通信消耗.提出一种基于位置感知和语义的P2P网络模型,在覆盖网络的构建过程中考虑底层的物理网络拓扑,同时兼顾拓扑结构与共享资源之间的相关性.该网格模型基础上,给出了语义搜索算法.实验结果表明,提出的网络模型能够减少覆盖网络与底层物理网络拓扑不匹配带来的网络开销,并且在查找性能和网络开销之间取得了良好的平衡.Unstructured P2P networks are widely used and popular due to their small maintenance overhead.However these systems suffer from the significant network overhead on flooding-based search.This paper proposes a Semantic and Location-aware based Unstructured P2P network（SLUP） taking advantage of the heterogeneous characters of nodes to cluster the nodes according to similar shared resources and physical distance.SLUP can not only alleviate the topology mismatch problem between the P2P overlay networks and the physical network,but also construct the relationship between resources and topology to increase the chances of finding the matching resources.Furthermore,semantic-based search algorithm is presented.Comparing the search performance in various overlay networks,simulation experiments show that SLUP and semantic-based search algorithm can significantly shorten the latency of the searching process and reduce the searching overhead.

关 键 词：P2P 覆盖网络 位置感知 资源搜索 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]