| 本文已被：浏览 2439次 下载 5266次 |
码上扫一扫！ |
|
|
| 一种基于共享执行策略的间隔查询优化技术 |
|
周新1,2, 张孝1,2, 薛忠斌1,3, 王珊1,2
|
|
1.数据工程与知识工程教育部重点实验室(中国人民大学), 北京 100872;2.中国人民大学 信息学院, 北京 100872;3.神华国华(北京)电力研究院有限公司, 北京 100069
|
|
| 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. |
| 关键词: 间隔查询 优化 共享执行 |
| DOI：10.13328/j.cnki.jos.005013 |
| 分类号: |
| 基金项目: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） |
|
| Technique Based on Shared Execution Strategy for Optimizing Interval Query |
|
ZHOU Xin1,2, ZHANG Xiao1,2, XUE Zhong-Bin1,3, WANG Shan1,2
|
|
1.Key Laboratory of Data Engineering and Knowledge Engineering of the Ministry of Education(Renmin University of China), Beijing 100872, China;2.School of Information, Renmin University of China, Beijing 100872, China;3.Guohua(Beijing) Electric Power Research Institute Co. Ltd., Beijing 100069, China
|
| Abstract: |
| As an important query type, interval query is widely used in social networks, information retrieval and database domain. Many kinds of optimization techniques have sprung up to support effective interval query. Although existing methods are efficient to handle single query, they all suffer from performance problem when the concurrent query loads exceed the processing capacity of the server such that more than 70% queries couldn't receive the results in the expected time. To solve this problem, this paper presents a method named SESIQ (shared execution strategy for interval queries). SESIQ batches interval queries, analyzes common operations among a group of interval queries and reduces duplicate data access to lower the cost of disk I/O and network transmission. The paper theoretically studies and analyzes SESIQ, and demonstrates the feasibility by large number of experiments based on two types of real datasets. Results show that SESIQ improves the performance of interval query by several ten folds. |
| Key words: interval query optimization shared execution |
