首页> 外文会议>Workshop on Intelligent Systems and Applications >An Effective Search Scheme for Gnutella-Like P2P Networks
【24h】

An Effective Search Scheme for Gnutella-Like P2P Networks

机译:GNutella样P2P网络的有效搜索方案

获取原文

摘要

In Gnutella-like P2P networks, blind flooding causes a large number of redundant messages, and this makes such networks not scalable. Previous researchers noticed three types of nodes in P2P networks: super nodes which share a large number of resources and tend to have long online time, stable nodes sharing a number of resources and having long online time, and dynamic nodes that share few resources and have short online time. Then they utilize an analytical model and simulation to obtain the result that, to keep search efficiency of P2P systems high, the resources of super and stable nodes can serve queries in large areas and the resources of dynamic nodes can only serve queries in local areas. Based on the above result, we present a new Differentiated Publishing (Diff-Publishing) architecture to make search efficiency of Gnutella-Like P2P systems high. The introduced network is a two-tier topology, neighbor overlay aims to let the resources of dynamic nodes serve queries in local areas, and semantic overlay tries to congregate the semantic relevant stable and super nodes to answer queries in large areas. Given a query, Diff-Publishing first routs it to the query generator's neighbors, and if it fails to hit the targets, it will be forwarded to the relevant semantic overlay. Simulation results show that the proposed search scheme greatly improves search efficiency of P2P networks while introducing only modest maintenance cost.
机译:在Gnutella样P2P网络中,盲目泛洪导致大量的冗余消息,这使得这种网络不可扩展。以前的研究人员注意到P2P网络中的三种类型的节点:超级节点,它共享大量资源,往往具有长长的在线时间,稳定的节点共享许多资源,并且具有长期的在线时间,以及共享资源的长期的节点,并且具有少量资源并具有简短的在线时间。然后,它们利用分析模型和仿真来获得结果,以保持P2P系统的搜索效率高,超级和稳定节点的资源可以在大面积中提供查询,动态节点的资源只能在本地区域的查询中服务。基于上述结果,我们提出了一种新的差异化发布(Diff-Publishing)架构,以提高GNutella样P2P系统的搜索效率。介绍的网络是一个双层拓扑,邻居叠加旨在让动态节点的资源在本地区域的查询,并且语义覆盖试图聚集语义相关的稳定和超级节点,以在大面积中回答查询。鉴于查询,Diff-Publishing首先将其路由到查询生成器的邻居,如果无法达到目标,则将转发到相关的语义覆盖。仿真结果表明,该拟议的搜索方案极大地提高了P2P网络的搜索效率,同时仅引入了适度的维护成本。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号