【摘要】 在全分布无结构P2P中,节点通常组织成为覆盖网络,通过查询消息在网络中广泛转发实现盲目搜索。由于数据存放位置独立于数据内容,一个节点并不清楚哪些节点更容易命中查询,因此发现路由方向感,提高查询消息转发有效性,对全分布无结构P2P搜索具有重要意义。在相关工作中,主要从用户兴趣、本体论等语义角度聚类用户,减小搜索范围。但当前语义获取和语义描述等工作还不甚成熟,因此这些方法并没有得到广泛采用。提出了一种以访问频率为路由方向感的新型搜索方法QRRO。在QRRO中,每个节点被分配一权重标识;节点仅仅为访问频率与节点权重接近的数据建立索引;基于访问频率建立存储内容和存储位置之间的耦合关系,形成路由方向感。模拟实验表明,QRRO在提高搜索成功率、降低搜索路径长度方面是有效的。而且,由于访问频率是每个文件都具有的非语义属性,因此QRRO具有通用性。更多还原

【Abstract】 In the Decentralized Unstructured Peer-to-Peer(P2P),the peers are usually organized to form an ad hoc overlay network,and the queries are propagated among the overlay to search blindly.Since the storage location is indepen-dent of the data content,a peer has no idea of which peer is more likely to satisfy a request.Therefore,it’s vital to find the routing orientation and improve the routing effectiveness.Some semantic methods,such as interest and Ontology,are commonly used in the related work to cluster the peers and to decrease the search range.However,these approaches have not been adopted widely because they are generally limited by the current immature semantic obtainment and description.This paper proposed a novel search method QRRO(Routing Orientation of Query Rate) with the Routing Orientation of Query Rate.In QRRO,each peer is allocated a weight identifier;a peer only indexes for the files whose query rate is close to its weight.Therefore,the coupling relation was built up and the routing orientation was formed through query rate.Our simulations show that QRRO is effective in improving the success rate and decreasing the search path length.What’s more,QRRO is a pervasive method because the query rate is a non-semantic property of each file.更多还原