面向应用服务的云规模虚似机性能监控与负载分析技术研究

发布时间：2018-10-05 16:35

【摘要】：云计算是一种新的分布式应用系统技术,旨在整合企业数据中心内部或多个数据中心间的资源,以服务的方式通过互联网向用户提供基础设施即服务(IaaS),平台即服务(PaaS)或软件即服务(SaaS)。云环境下运行着大量提供服务的虚拟机,虚拟机的性能监控与负载分析评估则是云环境下重要的组成部分,它对提高云服务质量起到重要的作用。并且云环境下提供应用服务的虚拟机通常会表现出丰富的负载特性,这些负载特性可以用来标识此应用服务的服务类型。因此,研究云规模虚拟机性能的监控与负载分析评估具有重要的意义。本文深入分析云计算与虚拟化的基本原理与体系结构、IaaS平台的通用架构模型。在OpenStack基础上设计并实现了IaaS平台的虚拟机性能评估方案,达到对虚拟机性能监控及负载分析评估的需求。针对云规模虚拟机负载信息难以获取的现状,设计针对虚拟机负载信息采集的数据采集模块,达到对虚拟机监控并高效的获取负载信息的目的。并提出了基于指定类型的优化的kmeans聚类算法对采集到的虚拟机负载信息进行处理,得出虚拟机的负载特征向量,用来描述些虚拟机的负载特性,负载特性分为：CPU密集型,内存密集型,网络密集型。为了让运行多种应用服务的虚拟机的用户和管理员对虚拟机的性能有更清晰的了解,通过公式运算出每种应用服务的基本特征值,为如何充分利用虚拟机与物理机的资源提供了一种解决思路。论文的最后对IaaS平台的虚拟机性能评估方案进行了功能与性能方面的测试,并对论文期间完成的工作成果进行了总结,并提出了下一步工作方向。
[Abstract]:Cloud computing is a new distributed application technology designed to integrate resources within or between enterprise data centers. Providing Infrastructure as a Service (IaaS), platform as a Service (PaaS) or Software as a Service (SaaS). Over the Internet A large number of virtual machines are running in the cloud environment, and the performance monitoring and load analysis and evaluation of virtual machines are important components in cloud environment, which plays an important role in improving the quality of service of the cloud. Moreover, virtual machines providing application services in cloud environments usually exhibit rich load characteristics, which can be used to identify the service types of this application service. Therefore, it is of great significance to study the performance monitoring and load analysis of cloud-scale virtual machines. This paper analyzes the basic principles and architecture of cloud computing and virtualization and the general architecture model of IaaS platform. The virtual machine performance evaluation scheme of IaaS platform is designed and implemented on the basis of OpenStack, which meets the requirements of virtual machine performance monitoring and load analysis and evaluation. Aiming at the current situation that the load information of cloud scale virtual machine is difficult to obtain, a data acquisition module is designed to collect the load information of virtual machine, so as to monitor the virtual machine and obtain the load information efficiently. An optimized kmeans clustering algorithm based on specified type is proposed to process the collected load information of virtual machine, and the load eigenvector of virtual machine is obtained, which is used to describe the load characteristics of some virtual machines, and the load characteristic is divided into two parts: CPU-intensive. Memory intensive, network intensive. In order for the users and administrators of virtual machines running multiple application services to have a clearer understanding of the performance of virtual machines, the basic characteristic values of each application service can be calculated by formula. It provides a solution for how to make full use of the resources of virtual machine and physical machine. At the end of the thesis, the function and performance of the virtual machine performance evaluation scheme of IaaS platform are tested, and the results of the work done during the thesis are summarized, and the next work direction is put forward.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP302

【相似文献】