一种非结构化P2P系统搜索算法的研究

Research of lookup algorithm on unstructured P2P system

P2P系统是目前计算机科学研究的热点领域,其搜索算法是该领域当前研究的重要问题之一,它直接关系到P2P系统的可用性.以往的非结构化的P2P系统采用的是无确定目标的自由搜索协议,它具有搜索效能低,无可扩展性的缺点.针对这些不足,文中提出了基于直接相邻优先和聚集度大优先策略的快速搜索算法,并设计实现了基于冗余扩散策略的资源索引建立算法.经对比试验证明,在相同情况下,采用文中所述的算法进行搜索比采用原有的洪泛算法搜索协议和索引算法进行搜索能够覆盖更多的节点,同时平均路径长度较小,算法具有良好的搜索性能.

Peer-to-peer （P2P） systems represent a current hotspot in the field of computer science research, espe cially their search algorithms, which directly relates to the usability of the P2P system. Unstructured P2P systems adopting a free search protocol have no certain target, resulting in low search capability and no scalability. To improve weak performance, a fast lookup algorithm is proposed based on direct adjacency first and larger connectivity first policies as well as a resource index building algorithm on redundant diffusing policies. Experiments show that the proposed algorithms can cover more nodes and have shorter average path length for the same instance than the initial P2P flooding routing protocol and the building of a resource index. The algorithms also exhibit satisfactory march performance.