如何快速准确搜索资源是DHT网络最核心的问题,在DHT之上建立逻辑的关键字搜索层是一个比较好的解决方案.逻辑层采取什么结构能更好地提高效率是一个值得研究的问题.现有的研究主要基于超立方体结构提出相应的索引和搜索算法,该方法当查询关键字数目较少时搜索效率很低.用改进的超立方体互连圈结构(MCCC)代替超立方体作为逻辑层来克服这一弱点.基于MCCC结构,提出了一个更高效的索引计划和搜索算法MCCCS.理论分析和实验结果证明,与基于超立方体的搜索算法相比,MCCCS搜索算法在用户提供的查询关键字较少时有更好的性能.
How to locate resources efficiently is a key issue in DHT-based network. A good solution is to build a logic keyword search layer on DHT. However, it's a challenging task to choose the proper structure for the logic layer for better searching efficiency. Existing techniques propose index scheme and search algorithm based on hypercube, the efficiency of which is low when there are few query keywords. The paper addressed the problem by replacing hypercube with a modified Cube-Connected-Cycle (MCCC). A better index and search scheme based on MCCC called MCCCS is proposed. It is demonstrated by experiments and analysis that the MCCCS search scheme works more efficiently under a MCCC when the number of query keywords is comparatively small.