摘要：

基于Kleinberg小世界模型设计一个结构化P2P网络文本检索协议。P2P网络由一些相互连接节点类构成，根据节点文本相似程度实现复杂查询。通过预先设定节点类参数，可以动态地调整节点类大小来平衡网络负载。节点类之间存在一些长程连接，缩短了查询路由步数。实验结果表明，随着网络规模的扩大，搜索开销呈对数平方曲线增长，维护开销呈线性增长，选择合适的节点类内部节点的数量，可以使得总开销最小。

关键词: 对等网络, 小世界现象, 文本检索, 覆盖网

Abstract:

A protocol is designed for text retrieval in structured Peer-to-Peer(P2P) network, which is based on Kleinberg small world model. Nodes are clustered to support complex query according to the similarity of nodes. The clustered nodes can adjust their sizes to balance the load of network. Routing path length is shortened by the long-link between node clusters. Experimental results show that search cost is increased with the curve of the square of logarithm and maintenance cost is linearly increased as the scale of network goes up, the total cost can be minimized by choosing proper cluster size.

Key words: Peer-to-Peer(P2P) network, small world phenomenon, text retrieval, overlay network

中图分类号: 

• TP311