论文部分内容阅读
top-k查询要求查找出最符合需求的前k个结果,是对等网络中的重要数据处理技术。该文研究了结构化对等网络中数据在各节点上垂直划分的精确top-k查询处理,在3通信回合的三阶段阈值(TPUT)算法基础上提出了4回合阈值算法4R-TPUT。它由下界估计、剪枝和结果查找3个阶段组成,通过在TPUT的下界估计阶段增加一个通信回合来获取更多的数据信息以得到更准确的top-k下界估计和剪枝阈值,从而减少查询处理过程中的数据访问和传输量。实验表明:4R-TPUT相比于TPUT较大幅度降低了数据传输量,减小了查询响应时间,是一种更高效的top-k查询算法。
The top-k query requires finding the best k results that best meet the requirements and is an important data processing technique in peer-to-peer networks. In this paper, the accurate top-k query processing of data partitioned vertically on nodes in structured peer-to-peer networks is studied. A 4-turn threshold algorithm 4R-TPUT is proposed based on the three-pass threshold (TPUT) algorithm of 3 communication rounds. It is composed of three phases, namely, lower bound estimation, pruning and result finding. It can reduce more by adding a communication round in the lower bound estimation stage of TPUT to get more data information to get more accurate top-k lower bound estimation and pruning threshold Query data access and throughput during processing. Experiments show that 4R-TPUT is a more efficient top-k search algorithm than TPUT, which greatly reduces the data transmission and reduces the query response time.