P2ST:基于带权搜索树的P2P搜索模型

P2ST:A Weighted Search Tree-based P2P Searching Model

针对非结构化P2P系统搜索效率低的问题,提出了一种基于K叉带权搜索树的P2P搜索模型P2ST。模型构建了服务于搜索的k叉带权树,节点按查询命中率大小在树中由上至下排列,命中率大且稳定的节点处于树的上层,搜索时可由此确定消息扩散的方向。采用缓存上层节点、建立搜索结果和发起节点索引、过热资源复制、为叶节点添加远程邻居等方法进一步提高搜索效率和平衡负载。分析和仿真结果表明,提出的模型能大量减少无效消息,具有较高的搜索效率,且维护搜索树的开销较小。

Improving search performance is an important issue in Peer-to-Peer （P2P）network systems.Although many policies are brought forward to address the issue, the question still exists. A searching model based on weighted search tree is proposed to improve search performance of unstructured P2P networks. A logical weighted k-tree is set up according to historical hit ratio, peers that have high probability to hit the query rise to higher layer of the tree, so that peers that unstable and have few hot resources are usually on low layer of the tree. Methods are also employed to increase efficiency of search, such as Hot peers caching, remote peer connections, source peer index and result index, and so on. Performance studies based on analyses and simulations are carried out, the results show that the proposed model can avoid a large amount of unnecessary messages, the overhead to maintain the tree is minimal, and outperforms related works in terms of search efficiency and search latency.