Study On Skyline Query Processing In Distributed Environment

With the application and development of Internet technology,the diversification of data acquisition methods has caused explosive growth of data on the Internet.How to deal with massive amounts of data and recommend information of interest to users has become one of the important topics in the field of database research.As an important means to solve multiple criteria decision-making problems,skyline query has important application value in financial markets,environmental monitoring and other fields.At present,a lot of results have been achieved in skyline query in centralized environment.And many variants of Skyline query have been proposed,among which dynamic skyline query is an important one.However,with the increase of data volume,the traditional skyline query algorithms have been unable to meet the query needs of large-scale data.Therefore,this paper deeply studied problems of skyline query and dynamic skyline query,and two efficient algorithms are proposed to solve related problems.The main contributions are summarized as follows.(1)One algorithm is proposed to process the skyline query problem in distributed environment(DSQ).First,a hierarchical grid index structure based on Z curve is designed to manage data.Then,an algorithm DSQ is proposed to complete the skyline query in a distributed environment.DSQ contains two phases.In the first phase,first,based oh the hierarchical grid index,a data block filtering algorithm is proposed to complete the batch filtering of the data in each node.Then,an index of dominace graph is proposed to complete the filtering of local data tuples.In the second phase,a distributed scheduling strategy is adopted to merge candidate sets.This cheduling strategy can assign all calculation to all data nodes without the occurrence of bottleneck nodes.Finally,the correctness and effectiveness of the DSQ algorithm are evaluated via a series of experiments.(2)Aiming at the problem of distributed dynamic skyline query,algorithm DDSQ is proposed.DDSQ contains two phases.In first phase,first,based on the B-tree index,a local basic scanning algorithm BSAB is proposed,which can quickly complete the calculation of the distributed dynamic skyline candidate set by reducing the scanning space.Then,based on the distribution histogram index,an optimized scanning algorithm OSAB is proposed.Compared with BSAB,OSAB algorithm further reduces the scanning space and improves the calculation efficiency.In second phase,adopt distributed scheduling strategy to complete the merge calculation of distributed dynamic skyline candidate sets.Finally,a series of comparative experiments were conducted to verify the effectiveness of the DDSQ algorithm.