一种基于共享执行策略的间隔查询优化技术

发布时间：2018-02-24 04:29

本文关键词： 间隔查询优化共享执行　出处：《软件学报》2016年12期 　论文类型：期刊论文

【摘要】：间隔查询作为重要的查询类型,广泛应用在社交网络、信息检索和数据库领域.为了支持高效的间隔查询,涌现出多种优化技术.尽管已有方法能够快速响应单个间隔查询,然而当查询负载超过服务器的处理能力时,70%的查询均不能在期望时间内得到响应.针对这一问题,提出采用共享执行策略优化间隔查询的方法 SESIQ(shared execution strategy for interval queries).SESIQ对间隔查询进行批处理,分析一组间隔查询间可共享的操作,减少重复数据的访问,从而降低磁盘I/O和网络传输代价,提高检索性能.理论分析并实验验证了SESIQ的可行性,基于两种真实数据集的大量实验结果表明,SESIQ是有效的,间隔查询的检索性能可提升数十倍.
[Abstract]:Interval queries as an important query type, widely used in the social network, information retrieval and database fields. In order to support efficient query interval, many optimization techniques. Although the existing methods can quickly respond to a single query interval, however, when the query load exceeds the processing capacity of the service server, the 70% query cannot obtained in response expected time. To solve this problem, the implementation of sharing the optimal strategy of SESIQ interval queries (shared execution strategy for interval queries.SESIQ) for interval queries for batch processing, analysis of a set of interval queries can be shared between operations, reduce duplication of data access, so as to reduce disk I/O and network transmission cost, improve retrieval performance. Theoretical analysis and experimental results demonstrate the feasibility of SESIQ, two kinds of real data set a large number of experiments show that SESIQ is effective, The retrieval performance of spaced queries can be increased dozens of times.

【作者单位】：数据工程与知识工程教育部重点实验室(中国人民大学);中国人民大学信息学院;神华国华(北京)电力研究院有限公司;
【基金】：国家自然科学基金(61432006) 中国人民大学科学研究基金(中央高校基本科研业务费专项资金)(10XNI018)~~
【分类号】：TP311.13

【相似文献】