基于OpenStack的虚拟机集群监控系统的设计与实现

发布时间：2018-06-17 13:23

本文选题：云计算 + 物理机监控　；参考：《北京邮电大学》2014年硕士论文

【摘要】：随着云计算的不断发展,工业界对云计算的重视程度不断增加,几年时间内,云计算已经成为了最热门的技术之一。云计算将各种软硬件结合起来,构成一个计算能力强大的平台,通过网络向用户提供各种各样的服务,满足用户的需求。在这样的环境下,OpenStaCk开源云平台出现了。OpenStack可以帮助企业快速构建云平台,向用户提供云服务,但是OpenStack自身并不完善,尤其是对物理机和虚拟机的监控。目前,OpenStack不支持对物理机性能的监控,用户自己设计调度算法时,不能有效获得物理机相关信息,并且对虚拟机的监控局限于启动、暂停等状态,不能获得虚拟机运行状态,根据这些参数进行扩容。基于上述问题,本课题对虚拟机集群监控系统进行研究与实现。本文在教育部中国移动科研基金2012年度项目“面向互联网的业务支撑系统关键技术及方案研究”课题中关于中国移动私有云资源池系统研究的基础上,完成了针对OpenStack云平台的虚拟机集群监控系统的设计与实现。系统为OpenStack云平台提供监控服务,对物理机和虚拟机进行监控,保证OpenStack各个服务的正常运行,为用户提供一个可靠、稳定的云服务平台。本文首先介绍云计算及云平台的概念、发展现状,对现有的各种各样的云平台进行对比研究。其次对云平台的关键技术虚拟化技术进行研究,分析了几种常见的监控技术和Libvirt。然后,结合实验室环境与中国移动私有云资源池系统,从结构、业务、监控功能端三个方面进行需求分析。在此基础上,提出了一种对OpenStack云平台下物理机和虚拟机监控的方案。监控代理进行物理机和虚拟机监控性能数据的采集,将采集的数据发送至监控服务端,监控服务端将数据写入数据库,监控功能端提供查询和报警等功能。本文从监控代理、监控服务端、监控功能端这几个方面对监控系统进行了设计和实现。本文提出了一种针对OpenStack云平台进行监控的方案。由于OpenStack云平台本身并不具有对物理机和虚拟机进行监控的能力,本文能够很好的弥补其不足,同时,本文提出的方案将监控性能数据存储,然后进行定制化的需求开发,对监控系统或者OpenStack云平台的功能进行扩展。
[Abstract]:With the development of cloud computing, industry pay more and more attention to cloud computing. In a few years, cloud computing has become one of the most popular technologies. Cloud computing combines all kinds of software and hardware to form a powerful computing platform. It provides users with various services through the network to meet the needs of users. In such an environment, OpenStaCk open source cloud platform has appeared. OpenStack can help enterprises quickly build cloud platform and provide cloud services to users, but OpenStack itself is not perfect, especially the monitoring of physical machines and virtual machines. At present, OpenStack does not support monitoring the performance of physical machines. When users design scheduling algorithms, they can not effectively obtain relevant information of physical machines, and the monitoring of virtual machines is limited to startup, pause and other states, and can not obtain the running state of virtual machines. The capacity is expanded according to these parameters. Based on the above problems, the virtual machine cluster monitoring system is studied and implemented in this paper. This paper is based on the research of China Mobile Private Cloud Resource Pool system in the project of "key Technologies and schemes of Internet-Oriented Business support system" of the China Mobile Research Foundation of the Ministry of Education in 2012. The virtual machine cluster monitoring system for OpenStack cloud platform is designed and implemented. The system provides monitoring services for OpenStack cloud platform, monitors physical computers and virtual machines, ensures the normal operation of OpenStack services, and provides a reliable and stable cloud service platform for users. This paper first introduces the concept and development of cloud computing and cloud platform, and makes a comparative study of all kinds of cloud platforms. Secondly, the key technology of cloud platform virtualization is studied, and several common monitoring technologies and Libvirtare analyzed. Then, combining the laboratory environment with the private cloud resource pool system of China Mobile, the requirements are analyzed from three aspects: structure, service and monitoring function. On this basis, a scheme to monitor physical machines and virtual machines on OpenStack cloud platform is proposed. The monitor agent collects the performance data of the physical computer and virtual machine, sends the collected data to the monitoring server, writes the data to the database, and provides the functions of query and alarm. This paper designs and implements the monitoring system from the aspects of monitoring agent, monitoring service and monitoring function. This paper presents a scheme for monitoring OpenStack cloud platform. Because OpenStack cloud platform itself does not have the ability to monitor physical machines and virtual machines, this paper can make up for its shortcomings. At the same time, the scheme proposed in this paper will monitor the performance data storage, and then develop the customized requirements. Extend the functionality of the monitoring system or OpenStack cloud platform.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06;TP311.52

【参考文献】