超级节点网络中并行查询和合并机制研究

Research on parallel query and merging mechanism in super-peers networks

针对超级网络中超级节点可能会导致网络瓶颈、检索结果重复问题,提出一种并行查询和合并机制。首先提出超级节点选择查询节点的算法,以减少超级节点的存储和计算负担。然后提出选择下载节点的合并算法,以得到高质量、新颖和更能有效获取的检索结果。实验结果表明了该机制的有效性。

For a super-peers network facing a few challenges such as network bottleneck and retrieval result duplicated,a mechanism to parallel query and merge the results retrieved is proposed.Firstly,in order to decrease super-peer burden,a method of selecting query-peer in a super-peer is presented.The method and algorithm can let a super-peer connect more client-peers in the scope of load capacity.A method of detecting duplicated results is presented,and a strategy of selecting down-load peer for duplicated results based on download data quantity and response time.The strategy can efficient reduce retrieval overlap,network transfers quantity for downloading results and response time for acquiring results.The experimental results show that the proposed mechanism is efficient.