关键词: 对等网络, 导向机制, 统计导向表, 语义分析

Abstract: Unstructured peer-to-peer systems have inherent deficiencies， the biggest one is the tradeoff between the success rate and the traffic cost of data query. The traditional flooding-based query algorithm brings too much network overhead while the low-cost query algorithm， such as random walks and expanding ring， cannot guarantee the query success rate. This paper proposes a query guided mechanism by the combination of the statistic result of query history and semantic analysis. The statistic result of query history helps lift query efficiency through forwarding the queries to neighboring nodes. Simulation results show this mechanism can effectively increase the query success rate， and reduce the traffic cost and query latency.

Key words: peer-to-peer network, guided search, statistics guided table, semantic analyses