OpenStack云平台下的虚拟机监控与控制的研究与实现

发布时间：2018-01-14 09:21

本文关键词：OpenStack云平台下的虚拟机监控与控制的研究与实现　出处：《电子科技大学》2013年硕士论文　论文类型：学位论文

【摘要】：云计算作为一种技术，，正越来越多的受到企业和政府的重视。在短短的几年里，高速发展的云计算已经成为最热门的技术之一。它帮助企业节约经济成本，帮助用户更好的利用网络资源，满足了用户的日常生活的需要。它把各种成本较低的硬件，软件资源结合起来，组成一个整体的具有强大计算能力的平台，为用户提供各种计算资源和服务。从谷歌2003年发表的核心论文开始，到2006年亚马逊把云计算技术的商业化，再到ATT提供的主机托管服务，云计算发展为从内部IT系统到提供公共服务，从大量浪费资源到节约成本的工具。OpenStack云计算平台就是在这样的环境下产生起来的。本文首先阐述云计算的概念、发展，然后讲解了几种常用的云计算平台架构及其特点。接下来介绍了虚拟化技术，分析了几种常用的监控技术。结合虚拟机监控和控制的特点，分析了监控系统所使用的分布式数据库系统Cassandra和数据传输、通信的JSON格式数据以及虚拟机监控和管理的libvirt的特点。然后对虚拟机的监控和控制的需求进行分析，提出了一种对OpenStack云平台下的虚拟机监控的方案。通过监控代理，获取监控的虚拟机的性能信息，然后发送给监控服务器。监控服务器接收到监控的信息后，把监控的信息录入数据库。用户可以通过监控的界面或者API查询到监控的数据信息。监控代理采用心跳机制，每隔一段时间向监控服务器发送监控数据。关于虚拟机实例的控制，首先研究了云环境下虚拟机操作的流程特点，设计和实现了Portal Server来控制虚拟机实例。对于用户提交的虚拟机控制的请求，首先要通过VM的管理服务，接受用户的请求，然后通过统一的接口，来与OpenStack进行交互来操作虚拟机实例，从而实现对它的控制。本论文相关的先进性工作主要体现在以下的两个方面：（1）运用消息传输机制来解决监控代理与数据库之间以及监控服务器和监控请求客户端之间一对多的消息传输来实现高效的监控数据的传输。（2）提出了一种针对OpenStack云环境下虚拟机监控的方案。由于OpenStack本身没有提供虚拟机监控的应用，本文能够很好的弥补其不足，而且提供良好的监控接口。
[Abstract]:Cloud computing is a technology that is more and more attention by enterprises and government. In a few short years, the rapid development of cloud computing has become one of the most popular technology. It helps enterprises to save the economic cost, the use of cyber source to help users better, to meet the needs of the user's daily life. It the low cost of hardware, software resources combine to form a whole with powerful computing platform, provide all kinds of computing resources and services for users. The core papers published from 2003 to 2006 to Google, Amazon cloud computing technology commercialization, to provide ATT hosting service, the development of cloud computing from the internal IT system to provide public services, is in such an environment from a lot of waste of resources to save the cost of the.OpenStack cloud computing platform.
This paper first explains the concept and development of cloud computing, and then explains the platform architecture and characteristics of several commonly used cloud computing. Then introduced virtualization technology, analyzes several commonly used monitoring technology. Combined with the characteristics of virtual machine monitoring and control, analysis of the monitoring system of the distributed database system and Cassandra data transmission, JSON the format of data communication and the characteristics of virtual machine monitoring and management of libvirt. Then the monitoring and control of the virtual machine needs analysis, put forward a virtual machine monitor on the OpenStack cloud platform under the scheme. By monitoring the performance of information acquisition agent, virtual machine monitor, and then sent to the monitoring server. The monitoring server receives the monitoring information, the monitoring information into the database. Users can monitor the interface or the API query to the monitoring data of information monitoring. The agent takes the heartbeat mechanism from time to time to send the monitoring data. The monitoring server control on virtual machine instances, first studied the characteristics of the process virtual machine operating under cloud environment, the design and implementation of Portal Server to control the virtual machine instance. The user submitted a virtual machine control request, first through VM management service, to accept the user's request, and then through the unified interface to interact with the OpenStack to operate the virtual machine instance, so as to realize the control of it.
The related advanced work in this paper is mainly reflected in the following two aspects:
(1) we use message transmission mechanism to solve one to many messages transmission between monitoring agent and database, as well as between monitoring server and monitoring client, so as to achieve efficient transmission of monitoring data.
(2) a scheme of virtual machine monitoring under OpenStack cloud environment is proposed. Because OpenStack itself does not provide the application of virtual machine monitoring, this paper can make up for its shortcomings and provide good monitoring interfaces.

【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP302

【引证文献】