基于虚拟机调度与调整的数据中心节能技术研究

发布时间：2018-07-15 10:27

【摘要】：云计算使得计算资源可以按需提供,给信息技术产业带来了革命性的变化。使得用户可以像使用水、电一样按需购买计算资源。以亚马逊为首的许多IT厂商纷纷推出自己的云计算服务,建立了包含成千上万计算节点的数据中心。与此同时,数据中心能耗问题也越来越引起人们的注意。大量的能耗不但带来了更多的运行成本,也会带来大量碳化合物的排放并导致严重的环境问题。据统计,数据中心在2010年的电力消耗占当年世界总电力消耗的1.1%到1.5%,并且这一比例还在不断增加。本文首先对数据中心节能技术进行了归纳总结,对现有的节能技术按照层次进行划分,并将本文的研究目标确定为数据中心层级基于虚拟机调度和调整的节能技术研究。接下来通过对数据中心虚拟机调度能耗问题进行建模,对于将多个虚拟机分配到物理机上执行的问题,提出了通过最小化忙处理时间降低能耗的模型。对于该NP难问题,提出了近似度为2的近似算法并进行了证明,并通过模拟实验验证了该节能算法的有效性。本文接下来对开源数据中心解决方案OpenStack进行节能技术研究,具体分析研究了OpenStack中与节能有关的虚拟机迁移、虚拟机自动扩展集群等虚拟机调整功能模块。指出了利用这些功能模块进行节能研究的方向。设计实现了基于虚拟机迁移的节能模块,针对虚拟机自动扩展集群中现有虚拟机数目调整策略节能方面的不足,设计了一种新的虚拟机数目调整策略并通过实验证明了该策略的节能有效性。
[Abstract]:Cloud computing makes computing resources available on demand and revolutionizes the information technology industry. Allows users to buy computing resources on demand just as they use water and electricity. Many IT vendors, led by Amazon, have launched their own cloud computing services and built data centers with thousands of computing nodes. At the same time, the problem of energy consumption in data center has attracted more and more attention. Large amount of energy consumption not only brings more operating costs, but also brings a lot of carbon emissions and serious environmental problems. According to statistics, the data center's electricity consumption in 2010 accounted for 1.1 percent to 1.5 percent of the world's total power consumption that year, and this proportion is still increasing. In this paper, the energy saving technology of data center is summarized, the existing energy saving technology is divided into different levels, and the research goal of this paper is to study the energy saving technology based on virtual machine scheduling and adjustment in the data center level. Then, by modeling the energy consumption of virtual machine scheduling in the data center, a model of reducing energy consumption by minimizing the busy processing time is proposed to solve the problem of allocating multiple virtual machines to physical machines. For the NP-hard problem, an approximate algorithm with degree of approximation 2 is proposed and proved, and the effectiveness of the proposed algorithm is verified by simulation experiments. In this paper, the open source data center solution OpenStack energy saving technology research, the specific analysis and study of OpenStack related to energy saving virtual machine migration, virtual machine automatic expansion cluster and other virtual machine adjustment function modules. The research direction of energy saving using these function modules is pointed out. The energy saving module based on virtual machine migration is designed and implemented. A new strategy for adjusting the number of virtual machines is designed, and the efficiency of the strategy is proved by experiments.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP308;TP302

【相似文献】