云平台资源调度模拟系统—基于用户需求的拟机调度算法设计与实现.docx

本科毕业设计（论文）题目：云平台资源调度模拟系统—基于用户需求的虚拟机调度算法设计与实现学 院 软件学院专 业 软件工程学生姓名 学生学号 指导教师 提交日期 年 月 日摘要随着虚拟化技术、互联网技术、Web服务技术的发展，各机构与企业对云计算资源的需求增加，因此云平台规模也迅速增大。随着云平台规模的发展，云平台的虚拟机数目将随物理主机数量的增加而增长，虚拟机初始化部署策略作为云平台资源管理的关键部分，其性能优劣对云平台的虚拟机管理非常重要。本文主要设计并验证动态需求与能耗均衡的虚拟机初始化部署方法，并协助课题小组成员完成针对当前云平台虚拟机部署和动态迁移的模拟系统的设计与调试。动态需求与能耗均衡的虚拟机初始化部署方法的核心内容包括：（1）基于层次分析法引入权重向量对虚拟机进行多属性加权以此来特征化虚拟机配置；（2）基于相对部署能力的改进降序最佳适应启发式放置方法；（3）低能耗优先的主机开启方法。本文针对该方法设计并实现了一个本地化运行的初始化部署的测试系统以及一个基于Web架构的远程初始化部署系统。为验证该方法的性能，本文将该方法与其他经典的方法（例如轮询方法，基于最佳适配改进的方法等）进行比较。仿真实验中构建不同场景并运行测试，通过综合比较各场景运行结果的平台总能耗、平台资源损耗值、用户需求违背率等评价指标，得出如下结论：该方法保证用户需求的同时，能有效的降低云平台的能耗。尤其在异构资源环境下和用户对虚拟资源的异构需求下，优势更明显。关键词：云计算；虚拟机初始化部署；动态需求；多属性加权；能耗AbstractAs virtualization technology, the rapid development of Internet technology, Web services, technology, institutions and enterprises demand for efficient, flexible, super computing resources more and more, while the initial deployment aspects of the problem such as loss of energy consumption is becoming more and more important. Main purpose of this paper is to validate the virtual machine to initialize the deployment of the performance of the algorithm and build a simulation system of virtual machine initialization deployment algorithm.In this process, to show the user the initial deployment is a dynamically based on the needs of multiple virtual machine resources of attribute weighted adaptive algorithm model of management framework, and combining with the algorithm of the model to implement the corresponding systems, and the performance of the algorithm is compared with the classical algorithm analysis, and the analysis of the corresponding conclusions.This paper mainly describes a dynamically based on the needs of multiple attribute weighted adaptive virtual machine resources management framework, the framework will contain the relevant business model, and the concrete implementation steps of the algorithm. At the same time we will also point out that the algorithm performance evaluation