基于OpenStack的资源负载预测方法研究  

Study onload prediction method based on OpenStack

作　　者：李丹程[1] 王晓晨[1] 宋晓雪[1] 马东琳[1] 李全祚 

机构地区：[1]东北大学软件学院,沈阳110819

出　　处：《计算机应用研究》2014年第7期2178-2182,共5页Application Research of Computers

基　　金：辽宁省科技项目博士启动基金资助项目(20111001);中央高校基本科研业务费资助项目(N110417004);辽宁省科技攻关项目(2011216027)

摘　　要：数据中心的能耗在云计算高速发展的同时不断增加,提高服务器资源利用率,进而提高数据中心的能效成为新的研究方向。基于OpenStack开源云计算平台,利用回归分析方法,预测云计算环境下服务器负载情况的变化,提出资源动态调度算法,将计算任务迁移到更合适的服务器,达到提高集群计算能效的目的。通过CPU计算能力的实验验证了该算法的有效性,实现了云计算环境下CPU计算能效的提高。The energy consumption of data center is continuous increasing with the development of cloud computing. How to improve the resource utilization of the server,and boost the energy efficiency of data center,becomes a new research area. This paper proposed a dynamic resource scheduling algorithm based on OpenStack cloud computing platform,utilized the analysis of regression to predict the variation of server load,and migrated computational tasks to more appropriate servers to improve the efficiency of clustercomputing. Through the experiment of CPU computational capabilities proves the effectiveness of the method and achieves an improvement of CPU computational efficiency in the circumstance of cloud computing.

关 键 词：云计算 OPENSTACK 回归分析 负载预测 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]