基于数据中心的能耗测试有效查询优化技术研究  

Research on Effective Query Optimization Technology Based on Data Center Energy Consumption Eest

作　　者：葛警军 王更 邢金芳 Ge Jingjun;Wang Geng;Xing Jinfang(SIPPR Engineering Group Co.,Ltd.,Zhengzhou,China)

机构地区：[1]机械工业第六设计研究院有限公司,河南郑州

出　　处：《科学技术创新》2023年第24期84-87,共4页Scientific and Technological Innovation

摘　　要：目前对于大规模数据中心最重要的就是降低能耗开销,建设绿色数据中心,在对需求满足的前提下如何使数据库系统节约能耗,是目前重点的研究课题。能耗有效性就是使用电能更少,提供的服务相同,能耗有效性越高,说明数据库系统能使用更少的能耗达到相同的性能。因此,提供一种面向绿色数据中心的能耗有效查询优化方法,该方法通过建立功耗预测模型,来对执行中的平均功能进行准确的预测,在PostgreSQL查询优化器中结合预测时间成本与能耗成本的新的查询执行代价模型,构件数据库系统能耗测试平台,基于TPC-H与TPC-C基准在PostgreSQL上进行实验。结果:本文提出的功耗预测模型的准确度相比已有方法更高,另外,性能退化度因子能为能耗和性能之间提供折中的方案,且性能退化度因子合理的设置,其能耗有效性比原始PostgreSQL更高。At present,the most important thing for large-scale data centers is to reduce energy consumption costs,build green data centers,and how to save energy consumption in database systems under the premise of meeting demand is a key research topic at present.Energy efficiency means that less energy is used,the same services are provided,and the higher the energy efficiency,the database system can use less energy consumption to achieve the same performance.Therefore,this paper provides an effective query optimization method for energy consumption for green data centers,which accurately predicts the average function in execution by establishing a power consumption prediction model,combines a new query execution cost model that predicts time cost and energy consumption cost in PostgreSQL query optimizer,builds a database system energy consumption test platform,and experiments are carried out on PostgreSQL based on TPC-H and TPC-C benchmarks.Results:The accuracy of the proposed power prediction model is higher than that of the existing methods,in addition,the performance degradation factor can provide a compromise scheme between energy consumption and performance,and the performance degradation factor is set reasonably,and its energy consumption effectiveness is higher than that of the original PostgreSQL.

关 键 词：功耗模型 代价模型 查询优化器 能耗有效性 数据中心 

分 类 号：TP308[自动化与计算机技术—计算机系统结构]