2017,53(15)
1 引言
云计算作为一种高可扩展性和高可用性的商业计
算模式,越来越被学术界和工业界熟知。当前实现低成
本、高效、安全、易用的云计算系统面临诸多挑战,其中
基于威布尔分布的云计算能耗优化模型研究
郭 栋
1
,肖清泰
1,2
,徐建新
1,2
GUO Dong
1
, XIAO Qingtai
1,2
, XU Jianxin
1,2
1. 昆明理工大学 质量发展研究院,昆明 650093
2. 冶金节能减排教育部工程研究中心,昆明 650093
1.Qual ity Development Institute, Kunm ing University o f S cience and Technology, Kunming 650093, China
2.Metallurgical Engineering Research Center of the Ministry of Education for Energy Conservation and Emissions Reduction,
Kunming 650093, Chi na
GUO Dong, XIAO Qingtai, XU J ianxin. Cl oud computing energy co nsu mption optimization model res earch based
on Wei bull distribution. Computer Engi neering and Ap plic ations, 2017, 53(15):77-84.
Ab stract: In the cloud computing environment, due to complex state of virtual machine normal operation, a novel m ethod,
processor running st ate is thought as har dware/software serie s repairable system and norma l operation is described by
Weibull distribution, is proposed to remit serious energy problems. The correlation model among processor util ization,
energy a nd given task completion time is structured by a semi-Markov process which for mulated for virtual machine
operation, Laplace-Stielties transform which used to simplify numeric al calculation, a nd Bayesian theorem which used to
remove restricted cond ition. Combined with repairable s ystem lifetime distribution theory, probability dens ity function of
VMs normal operation is changed. I t can obtain processor operati on energy consumption and a give n task completion
tim e under different shape parameter of Weibull distribut ion. The implied relationship among shape parameter, processor
u tilization, energy and performance is analyzed to give some rat ional proposal in order to e ffectively redu ce energy
consumption. Numerical examples results show that increasing shape parameter is superior to improving processor utilization
about reducing energy consumption. Increasing shape parameter by improving processor configuration not only can signif-
icantly reduce energy consumption , and can avoid depletion of system perf ormance.
Key words: cloud computing; energy consumption; Weibull distribution; shape parameter; semi-Markov proce ss
摘 要:针对云计算环境中复杂的虚拟机正常运行状态,提出将虚拟机运行看成由硬软件串联的可修复系统,用威
布尔分布描述虚拟机硬软件正常运行的方法来建模缓解云计算高能耗问题。运用半马尔科夫过程刻画虚拟机运行
过程、Laplace-Stielties 变换简化数值计算、Ba yesian 定理去除限制条件,构建处理器利用率与能耗、性能的关系模
型。再结合可修复系统寿命分布理论,调整虚拟机正常运行的威布尔分布函数,得到不同形状参数下处理器运行能
耗以及给定任务完成时间,最终分析形状参数、处理器利用率与能耗-性能隐含关系并给出有效降低处理器运行能
耗的合理化建议。数值分析表明:增大形状参数比提高利用率更显著降低处理器运行能耗 ;优化虚拟机配置使得形
状参数变大,可以明显降低处理器运行能耗,同时避免云系统性能过度损耗。
关键词:云计算;能耗;威布尔分布;形状参数 ;半马尔科夫过程
文献标志码:A 中图分类号:TP316 doi:10.3778/j.issn.1002-8331.1607- 036 8
基金项目:国家自然科学基金(No.51406071,No.51666006)。
作者简介:郭栋(19 90—),男,硕士,主要研究领域为大数据与云计算,E-mail:7 51517290@qq.com;肖清泰(1989—),男,博士,主
要研究领域为节能减排;徐建新(1983—),男,通讯作者,博士,副教授,主要研究领域为数据挖掘。
收稿日期:2016-07-21 修回日期:2016-11-22 文章编号:1002-8331(2017)15-0077-08
CN KI 网络优先出版:2017-03-04,http://kns.cnki.net/kcms/detail/11.2127 .TP.20170304.1810 .034.html
Computer Engineering and Applications 计算机工程与应用
77