一种基于模糊聚类的资源发现策略

Chord是一款较为成功的P2P资源发现算法，但在网格环境中由于网格资源固有的特性，影响了其发现效率。针对网格资源的复杂多样性以及很难精确分析其相似程度，提出利用模糊聚类的方法对资源进行属性划分，将资源分成多个资源相似集合，结合Chord路由算法，相似资源存放在较近的Peer上组织成2维Chord结构。实验结果表明，与原Chord路由算法相比，该策略能有效降低查寻时间，资源发现效率明显提高。

Approach of resource discovery based on fuzzy clustering

Chord is a successful P2P resource discovery algorithm, but the inherent characteristic of grid resource affects the efficiency of discovery in grid environment. Focusing on the complexity and diversity of grid resource, and the fact that its similar degree was hardly analyzed in grid computing, a model for resource discovery based on fuzzy clustering was presented in this paper. This method will partition grid resource into many similar sets according to their properties. Similar resources are stored in the near Peer to be organized into 2D Chord to reduce access latency and raise lookup efficiency. Experimental results show that this approach can effectively reduce access latency compared with the original Chord. Thus, the efficiency of resource discovery is apparently improved.