摘要
提出了一个对等网络中基于"簇"的搜索机制,利用信息社会中数据的存放具有结构化的特点,设计了一个算法在具有相似"兴趣"的Peer之间建立邻居关系,形成一个"簇".转发查询请求时优先考虑"兴趣"更接近的簇内的Peer,大大提高了搜索效率,同时保持了非结构化系统的良好的容错性且支持关键字部分匹配.
This paper presented a 'cluster' based search scheme in peer-to-peer netowrk. Its idea is based on the fact that data distribution in information society has a structured feature. An algorithm was designed to cluster the peers that have similar interests. When receiving a query request, a peer is preferentially forwards to the peer that is in the same cluster and has more similarities. Through this scheme the search efficiency is remarkably improved, meanwhile a good resilience to peer failure is reserved. Partial-match queries are supported either.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2004年第1期91-95,共5页
Journal of Shanghai Jiaotong University
关键词
对等网络
簇
叠加
结构化
请求路径长度
peer-to-peer network
cluster
overlay
structured
request pathlength
