支持模糊搜索的结构化P2P算法

Structured P2P algorithm with fuzzy query

传统的结构化P2P系统不支持模糊搜索,针对结构化P2P系统的模糊搜索问题,提出一种改进的P2P算法。为应用系统建立类别树,使用源自类别树的类别向量而非传统的二进制码标识结点和实体,将实体信息存储在类别相近的结点,形成按类别聚集的结点和实体分布。在此基础上,提出路由算法和结点发现算法,实现了模糊搜索功能。实验证明,算法具有较高的查全率。算法已在某服务计算平台中成功应用,该平台已通过验收并连续运行近一年。

Aiming at solving fuzzy search problem existing in general structured peer-to-peer network, an improved＇algo- rithm is proposed in this paper. Items and nodes are identified by category vectors that come from a classification tree estab- lished for application system, instead of traditional binary code. hems are stored in the nodes with similar category, and gathered together according to the classification tree. On this basis, the routing algorithm and node discovery algorithm are discussed, and the function of fuzzy query is realized. The discovery algorithm can achieve high rate of recall, which has been proved by experiments. The approach has been successfully applied in a service computing platform, which has al- ready been verified and well operated nearly a year.