摘 要

随着云计算在全世界范围内的推广，越来越多的企业和大型公司开始接触云计算和云平台，可以说云计算和云平台已经成为了一种发展趋势，并且不断壮大，开始被各国政府重视，在大势所趋下，我国也逐渐开始重视云计算方面的发展，从“十二五”规划中就可以看出，国家对云计算方面的重视程度。

但是随着云计算的发展，越来越多的地方开始组建大型的数据中心，而数据中心的能耗十分庞大，这并不符合绿色IT的倡导，所以使数据中心能够在保证高效率高性能的前提下，节约能源，减少能量的消耗便成为了重中之重。由于我们不可能在大型的数据中心中进行对能耗的研究和实验，刚好CloudSim为学者们提供了一个模拟大型数据中心环境的平台，能够在CloudSim上进行算法的实验和研究，为解决大型数据中心的能耗问题带来一些解决措施。近几年，无论是国内还是国外的专家学者都提出了自己的很多资源调度算法，但是各有各的优缺点，所以在这一方面还是值得我们去研究一番。

首先要对CloudSim平台有一定的了解，学习CloudSim平台的体系结构和仿真流程，学会并掌握重新编译CloudSim源代码的方法，了解虚拟机分配起到关键作用的类，知道CloudSim平台是如何进行虚拟机的分配，并且对算法有一定的了解。

虚拟机的分配是服务器和数据中心能量消耗的中心，所以一个好的虚拟机分配算法能够明显的降低数据中心的能量消耗，比较古典的算法有先来先服务算法、首次适应算法、最佳适应算法和最坏适应算法等等。本文将讨论的是BFD算法，即降序的最佳适应算法。

通过在CloudSim平台进行仿真和模拟，利用自带的power实例来进行对算法的比较和分析，来得出BFD算法在虚拟机分配策略上的表现是本文的重点。在保证性能为前提的情况下，利用BFD算法进行对虚拟机的分配，使得模拟数据中心的消耗降低，从而应用在大型数据中心上，降低能耗，为推动绿色IT产业的发展做出贡献，也为虚拟机分配策略的研究提供了参考和借鉴。

关键词：云计算；CloudSim平台；能耗；算法

Abstract

With the spread of cloud computing throughout the world, more and more enterprises and large companies have begun to contact cloud computing and cloud platforms. It can be said that cloud computing and cloud platforms have become a trend of development, and are growing to be paid attention by governments. In the trend of the times, China has gradually begun to attach importance to cloud computing. From the "12th Five-Year plan", we can see that the state attaches importance to cloud computing. Basic on analysing and researching on social apps at home and abroad,this article then raises this app development that focusing on undergraduate students.This app are available for users to log in and register. The users are able to release their own messages and comment other users’ message, besides, the private chat can be engaged in this app among user to user. And the management of dormitory members can be done by this app, too.

However, with the development of cloud computing, more and more places have started to form large data centers, and the energy consumption of data centers is very large. This does not conform to the advocacy of green IT, so it is the most important thing to save energy and reduce energy consumption on the premise of ensuring high efficiency and high performance. Since it is impossible for us to carry out research and experiment on energy consumption in large data centers, just CloudSim provides a platform for scholars to simulate the environment of large data centers. It can carry out experiments and Research on the algorithm on the CloudSim, and bring some solutions to solve the problem of the energy consumption of large data centers. In recent years, experts and scholars both at home and abroad have put forward a lot of their own resource scheduling algorithms, but they have their own advantages and disadvantages, so it is worth studying in this respect.

First of all, we should have a certain understanding of the CloudSim platform, learn the architecture and simulation process of the CloudSim platform, learn and master the method of recompiling the CloudSim source code, understand the key role of the virtual machine allocation, know how the CloudSim platform is how to allocate the virtual machine, and have a certain understanding of the algorithm.

The allocation of virtual machines is the center of the energy consumption of the server and data center, so a good virtual machine allocation algorithm can obviously reduce the energy consumption of the data center. The classical algorithms have first service algorithm, first adaptive algorithm, optimal adaptive algorithm and worst adaptive algorithm, and so on. This paper will discuss the BFD algorithm, that is, the best adaptation algorithm for descending order.

Through the simulation and Simulation on the CloudSim platform, the comparison and analysis of the algorithm with the power instance is carried out, and the performance of the BFD algorithm in the virtual machine allocation strategy is the focus of this paper. In the condition of guaranteeing the performance, the BFD algorithm is used to allocate the virtual machine, which reduces the consumption of analog data center, so as to reduce energy consumption, contribute to the development of green IT industry, and provide reference for the research of virtual machine allocation strategy.

Key Words：cloud computing；cloudsim ；energy consumption；algorithm

目 录

摘 要 I

Abstract II

第1章 绪论 1

1.1 研究目的及意义 1

1.2 国内外研究现状 1

1.3 论文的研究内容 2

第2章 相关技术概述 3

2.1 开发工具及环境 3

2.1.1 Eclipse 3

2.1.2 CloudSim平台 3

2.2 相关技术介绍 4

2.2.1 CloudSim体系结构 4

2.2.2 CloudSim配置 4

第3章 基于BFD算法的分配策略研究 5

3.1 系统模型 6

3.2能耗模型····························································6

3.3问题描述····························································6

3.4 算法描述 7

3.4.1装箱问题························································7

3.4.2bfd算法描述·····················································7

3.4.3算法流程图······················································8

3.5power类解析 9

3.5.1vmallcationpolicy抽象类 9

3.5.2powervmallocationpolicyabstract抽象类···························9

3.5.3子类powervmallocationpolicysimple······························9

3.6算法实现··························································10

第4章 实验分析 10

4.1运行结果··························································10

4.2结果分析··························································11

第5章 总结和展望····································

致 谢 20

第1章 绪论

1.1 研究目的及意义

云计算（Cloud Computing）是基于互联网的相关服务增加使用和交互，云计算中的云是指网络、互联网的一种说法，通过互联网提供动态易拓展且是虚拟化的资源。现阶段一种广为接受的云计算的定义是：云计算是一种付费使用的模型，它为可配置的计算机资源共享池（包括网络、服务器、存储、服务、应用）提供可用、方便、按需的网络访问，这些资源可以很快的提供，并且易于管理。或者与服务供应商进行很少的交互。