虚拟化云计算平台的能耗管理  

Power Management of Virtualized Cloud Computing Platform

作　　者：叶可江[1] 吴朝晖[1] 姜晓红[1] 何钦铭[1] 

机构地区：[1]浙江大学计算机科学与技术学院,杭州310027

出　　处：《计算机学报》2012年第6期1262-1285,共24页Chinese Journal of Computers

基　　金：国家"八六三"高技术研究发展计划重大项目基金(2011AA01A207);国家自然科学基金项目(11071215);教育部-英特尔专项基金(MOE-INTEL-11-06)资助~~

摘　　要：数据中心的高能耗是一个亟待解决的问题.近年来,虚拟化技术和云计算模式快速发展起来,因其具有资源利用率高、管理灵活、可扩展性好等优点,未来的数据中心将广泛采用虚拟化技术和云计算技术.将传统的能耗管理技术与虚拟化技术相结合,为云计算数据中心的能耗管理问题提供了新的解决思路,是一个重要的研究方向.文中从能耗测量、能耗建模、能耗管理实现机制、能耗管理优化算法4个方面对虚拟化云计算平台能耗管理的最新研究成果进行了介绍.论文分析了虚拟化云计算平台面临的操作管理和能耗管理两方面的问题,指出了虚拟化云计算平台能耗监控与测量的难点;介绍了能耗监测步骤及能耗轮廓分析方法;提出了虚拟机系统的整体能耗模型及服务器整合和在线迁移两种关键技术本身的能耗模型;从虚拟化层和云平台层两个层次总结了目前能耗管理机制方面取得的进展;并对能耗管理算法进行分类、比较.最后对全文进行总结,提出了未来十个值得进一步研究的方向.The high energy consumption of data center is a serious problem. Recently, with the rapid development of virtualization technology and the emergence of cloud computing paradigm, a large number of future-generation data center will use virtualization technology and cloud computing technology, due to the benefits of high resource utilization, flexible management, and dynamic scalability. The integration of traditional power management technology and virtualization technology provides new solutions to solve the power management issue of cloud data center, which is an important research direction. This paper focuses on the virtualized cloud computing platform, surveys the latest research results from the perspectives of power measurement, power modeling, implementation mechanisms, and optimization algorithms. We investigate the challenges of operation and power management in virtualized cloud platform, and try to find out the difficulties of power monitoring and measurement technology. We introduce the steps of power measurement and power profiling technology, and try to build the power model for the whole virtual machine system, also the power model for the technique of server consolidation and live migration. We summarize the advances on power management mechanisms from both virtualization level and cloud level, and classify and compare the existing power management algorithms. Finally, we summarize the contents of this paper and propose ten possible research directions in the future.

关 键 词：虚拟化 云计算 能耗管理 绿色计算 在线迁移 服务器整合 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]