基于OpenStack云平台资源调度方法的研究

发布时间：2018-05-16 23:20

本文选题：云计算 + 虚拟机放置　；参考：《南京邮电大学》2017年硕士论文

【摘要】：云计算技术是新型的IT模式,从网格计算、并行计算与分布式计算发展而来。云计算是指用户可以按照自己的计算需求,通过网络以一种较为容易扩展的方式来获取所需的计算资源(如计算、存储和网络等)的服务。云计算通过虚拟化技术将底层物理硬件虚拟化,形成巨大的虚拟资源池,然后以动态可伸缩的服务形式将资源提供给用户。随着云计算用户的持续增加,云数据中心的规模也在不断扩大,如何对物理资源进行分配,选择合适的虚拟机资源调度方式来提高云数据中心性能、提高云应用的性能和降低云应用处理时间,成为如今云计算要改善和解决的主要问题。本文立足于OpenStack这个开源云平台,结合云计算的虚拟机调度算法,对云数据中心的资源调度问题进行深入的研究。主要工作内容如下:(1)对云计算的基本概念和相关特性进行阐述,分析了云计算的三种应用框架。阐述了虚拟化的相关概念和主流的虚拟化技术。分析OpenStack云平台,对OpenStack平台的组织结构和逻辑架构进行详细的阐述,主要对核心组件进行了分析。(2)提出了基于粒子群算法的虚拟机放置策略,该策略为解决应用基于复杂网络具有较高通信时延的问题,该策略考虑了CPU、内存、带宽和网络通信量四种因素,通过建立云环境内部时延模型,利用改进的粒子群算法来降低应用的时延,提高运行效率。(3)基于OpenStack平台的Nova组件,设计虚拟机调度模块,基于WFPSO算法,对虚拟机调度模块进行系统架构与功能模块的设计,并将其集成在OpenStack平台上,并实现该平台的部署。(4)将实现虚拟机调度模块的OpenStack平台进行功能性测试,测试该调度模块的可用性与正确性。同时对调度模块的WFPSO算法进行性能测试,采用CloudSim仿真平台对该算法进行大规模集群测试,验证其可以较大幅度的提升云环境中应用的运行效率。
[Abstract]:Cloud computing technology is a new IT model developed from grid computing, parallel computing and distributed computing. Cloud computing refers to the services that users can obtain computing resources (such as computing, storage, network, etc.) through the network in a way that is easy to expand according to their own computing requirements. Cloud computing virtualizes the underlying physical hardware through virtualization to form a huge pool of virtual resources and then provides resources to users in the form of dynamically scalable services. With the continuous increase of cloud computing users, the scale of cloud data center is also expanding. How to allocate physical resources, select the appropriate virtual machine resource scheduling method to improve the performance of cloud data center, Improving the performance of cloud applications and reducing the processing time of cloud applications have become the main problems that cloud computing needs to improve and solve. Based on the open source cloud platform OpenStack and the virtual machine scheduling algorithm of cloud computing, the resource scheduling problem of cloud data center is deeply studied in this paper. The main work is as follows: (1) the basic concepts and related features of cloud computing are expounded, and three application frameworks of cloud computing are analyzed. This paper expounds the related concepts of virtualization and the mainstream virtualization technology. The OpenStack cloud platform is analyzed, the organization structure and logic structure of OpenStack platform are described in detail, and the kernel components are analyzed. (2) the strategy of virtual machine placement based on particle swarm optimization is put forward. In order to solve the problem of high communication delay based on complex network, this strategy considers four factors: CPU, memory, bandwidth and network traffic, and establishes the internal delay model of cloud environment. The improved particle swarm optimization algorithm is used to reduce the application delay and improve the running efficiency. The Nova component based on OpenStack platform is designed and the virtual machine scheduling module is designed. Based on the WFPSO algorithm, the system architecture and function module are designed for the virtual machine scheduling module. It is integrated on the OpenStack platform, and the deployment of the platform. 4) the virtual machine scheduling module will be implemented on the OpenStack platform for functional testing, testing the availability and correctness of the scheduling module. At the same time, the performance of the WFPSO algorithm of the scheduling module is tested, and the large scale cluster test of the algorithm is carried out on the CloudSim simulation platform, which verifies that it can greatly improve the efficiency of the application in the cloud environment.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09

【参考文献】