超对等网络中的轮廓查询优化

轮廓查询是近年来信息服务领域的一个研究重点和热点.现有的三阶段算法TPAOSS (Three-Phase Algorithm for Optimizing Skyline Scalar)至少存在如下两个缺陷:(1)在TPAOSS算法的第3阶段中,当网络节点上的对象个数较多时,Bloom filter的长度将呈指数级增长,从而严重影响获取子空间重复值的效率以及占用内存空间的大小;(2)TPAOSS算法只考虑预处理阶段的时间代价,而没有考虑各网络节点进行局部或全局子空间轮廓查询计算的效率.为此,提出一种适合超对等网络(Super-Peer Architecture,SPA)的子空间轮廓查询方法EPSSQDN (Efficient Processing of Subspace Skyline Queries in Distributed Networks).EPSSQDN算法有效解决了TPAOSS算法的的两个主要性能问题,并且显著提高了SPA网络中的子空间轮廓查询处理的效率.此外,为了能够进一步降低子空间上轮廓查询的时间开销以及网络节点间的数据传输量,我们给出新颖且有效的优化策略.实验结果表明,EPSSQDN算法比TPAOSS算法更能够缩短SPA网络中子空间轮廓查询的时间开销.

Optimizing Skyline Queries in SPA Distributed Networks

Skyline query has recently received a lot of attention in information service community.The TPAOSS (Three-Phase Algorithm for Optimizing Skyline Scalar) algorithm has two performance drawbacks:(1) in the third phase of TPAOSS,as the number of objects on net nodes increases,the length of bloom filter will increases exponentially,which will seriously influence the efficiency of obtaining replicated values and the occupation size of memory;(2) the TPAOSS algorithm does not consider the computation efficiency of local or global subspace skyline queries in each net node.Motivated by these facts,we propose EPSSQDN (Efficient Processing of Subspace Skyline Queries in Distributed Networks),an algorithm for efficient processing of subspace skyline queries in SPA distributed networks.Moreover,in order to further reduce the computation cost of subspace skyline queries and decrease the volume of data transferred,we present an efficient optimized techniques.Furthermore,we present extensive experiments that demonstrate our method is more advantageous than the TPAOSS algorithm.