基于DEA的能耗感知虚拟机资源分配算法  

Energy-aware Virtual Machine Resource Allocation Algorithm

作　　者：陈小娇[1] 陈世平[1,2] 

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093 [2]上海理工大学信息化办公室,上海200093

出　　处：《小型微型计算机系统》2015年第1期167-171,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61170277)资助;上海市教委科研创新重点项目(12zz137)资助;上海市一流学科建设项目(S1201YLXK)资助;上海市研究生创新基金项目(JWCXSL1202)资助

摘　　要：针对云计算虚拟化中心高能耗问题,提出一种能耗感知的虚拟机资源分配算法.该算法建立能耗模型,利用DEA评估模型对虚拟机部署方案历史资源需求和能耗进行评估,将CPU,磁盘,内存和带宽等资源作为投入指标,CPU,内存和磁盘的利用率和能耗之比作为产出指标,在评估过程中获取虚拟机资源的最佳权重和同时达到最佳技术效率和最佳投入规模的虚拟机部署方案.资源分配算法按照DEA评估结果进行虚拟机部署,DEA有效的虚拟机根据评估前的方案进行配置,DEA弱有效的根据DEA前沿面阴影计算出最佳投入资源来进行配置.实验结果表明,该算法不仅降低虚拟机能耗,还提高了资源利用率,达到节能效果.To resolve the high energy consumption in virtualization Cloud Computing Center, an energy-aware virtual machine resource allocation algorithm is proposed. The algorithm established energy consumption model and through DEA evaluation model to access Historical resources requirement and energy consumption of Virtual Machine Deployment Scenarios. It put CPU, disk, memory and bandwidth resources as input indicators and the ratio of CPU, memory and disk utilization and energy consumption as output indica- tors. In the access process it achieves the optimal weight of virtual machine resources and the deployment scenarios which reach the best technical efficiency and the best investment scale. Resource allocation algorithm deploys virtual machine according to DEA evalu- ation result. Virtual machines DEA effective are deployed according to pre-assessment scheme and virtual machine and the resource al- located to virtual machine DEA weak effective are calculated according to DEA frontier shadow. The experimental results show that the algorithm is not only effective to reduce virtual machine energy consumption, but also improve the resource utilization of physical servers,as much as possible to save energy.

关 键 词：云计算 虚拟化 能耗 节能 DEA 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]