考虑虚拟机生命周期和延时的云数据中心的节能算法研究

发布时间：2018-07-09 22:29

本文选题：云计算 + 数据中心　；参考：《电子科技大学》2012年硕士论文

【摘要】：云计算时代。如何对数据中心的资源进行高(?)和经济的管理是一个具有挑战性的问题.数据中心消耗了大量的能源,同时带来了环境污染-因此设计合理的节能调度算法对数据中心的资源进行管理,降低数据中心的能耗、减少运营成本具有十分重要的意义. 为了降低数据中心的能耗,本文对物理服务器的能耗进行了大量实验和数据分析,得出了数据中心能耗的计算模型,在此基础上设计了离线延时节能和在线延时迁(?)节能两种调度算法.算法主要特点包括考虑了虚拟机的生命周期问题,考虑了(?)线和在线两种不同的应用场景,并且在调度过程中考虑了延时的因素.算法通过对虚拟机任务请求的合理分配和迁移使数据中心在调虚过程中开启的物理服务器的数量和时间尽量减少,最后达到节能的效果, 最后通过和其他的四种节能调度算法的对比,发现在数据中心总能耗,数据中心开启的物理机数量最大值.数据中心开启的物理机的总时间、数据中心拒绝的任务请求数量等多个参数的对比中。本文设计的两种节能调度算法都有明显的优势.在相同虚拟机任务请求的情况下这两种节能调度算法数据中心的总能耗最少,使用的物理机资源最少。同时拒绝的任务请求数量也最少。
[Abstract]:Cloud computing age. How to raise the resources of the data center (?) And economic management is a challenging issue. The data center consumes a lot of energy and brings environmental pollution at the same time. Therefore, it is very important to design a reasonable energy saving scheduling algorithm to manage the resources of the data center, reduce the energy consumption of the data center and reduce the operating cost. In order to reduce the energy consumption of the data center, this paper makes a lot of experiments and data analysis on the energy consumption of the physical server, obtains the calculation model of the energy consumption of the data center, and designs the off-line delay energy saving and the on-line delay transfer (?) Energy saving two scheduling algorithms. The main features of the algorithm include taking into account the life cycle of virtual machine and considering (?) There are two different application scenarios, line and online, and the delay factor is taken into account in the scheduling process. The algorithm minimizes the number and time of physical servers opened by the data center in the process of adjusting virtual machine by reasonably allocating and migrating the task requests of virtual machine. Finally, the effect of energy saving is achieved. Finally, by comparing with other four energy-saving scheduling algorithms, it is found that the maximum number of physical machines opened in the data center is the total energy consumption of the data center. Data center turned on the total time of the physical machine, data center refused the number of task requests and so on in the comparison of multiple parameters. The two energy-saving scheduling algorithms designed in this paper have obvious advantages. In the case of the same virtual machine task request, the data center of these two energy-saving scheduling algorithms has the least total energy consumption and the least physical machine resources. At the same time, the number of task requests refused is the least.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP308

【共引文献】