云环境下资源管理运行时优化技术研究

发布时间：2018-05-12 00:30

本文选题：云计算 + 虚拟化　；参考：《复旦大学》2013年硕士论文

【摘要】：近年来,云计算因其动态分配、弹性扩展、资源共享、按需使用按需付费等特点,吸引了越来越多的关注。云计算不仅改变了当今IT基础设施的架构模式,也改变了云服务使用者获取IT资源、以及云服务提供者管理和提供软件、技术和解决方案的方式。云计算在为企业和用户带来便利的同时,也对云平台管理者提出了许多挑战。对于IaaS云而言,云平台旨在努力满足用户的资源需求。除此之外,如何从数据中心的层面,更加合理的分配和使用资源、保证应用性能和可扩展性、优化整体的资源利用情况、并降低运维和管理成本,都是云管理者必须面对与亟待解决问题。本项研究关注IaaS私有云资源管理的运行时优化技术。由于在IaaS私有云环境下,云管理者通常具有掌控云资源的更大自主权和管理其上应用的更高控制权,从而存在更广的技术优化空间。本文首先从促进虚拟机资源更合理分配和提高资源利用率的角度,提出了应用负载感知的虚拟机资源配置优化方案,以及基于资源预留与资源共享相结合的资源分配方式,旨在为虚拟机配置更贴合其真实需求的资源数量。进而,为了应对云环境中应用负载的动态变化,提出了优化的多层应用动态资源调整方案,主要研究了评估何时需要资源调整、以及如何进行调整两方面。最后,整合前两步的研究内容,提出了云环境资源管理运行时优化的整体系统架构和模块划分,并且利用业界主流的基准应用RUBiS进行检验,获得了该系统方案正确性与有效性的验证。
[Abstract]:In recent years, cloud computing has attracted more and more attention because of its dynamic allocation, flexible expansion, resource sharing and on-demand payment. Cloud computing has changed not only the architecture of today's IT infrastructure, but also the way cloud service consumers access IT resources and how cloud service providers manage and deliver software, technology and solutions. Cloud computing not only brings convenience to enterprises and users, but also brings many challenges to cloud platform managers. For the IaaS cloud, the cloud platform is designed to meet the resource needs of users. In addition, how to allocate and use resources more reasonably from the level of data center, ensure application performance and scalability, optimize the overall utilization of resources, and reduce the cost of transportation and peacekeeping management, Are cloud managers must face and solve problems. This study focuses on runtime optimization techniques for IaaS private cloud resource management. In the private cloud environment of IaaS, cloud managers usually have more autonomy to control cloud resources and higher control power to manage their applications, so there is more space for technical optimization. In this paper, from the point of promoting the rational allocation of virtual machine resources and improving the utilization of resources, this paper proposes an optimal resource allocation scheme based on load awareness, and a resource allocation method based on the combination of resource reservation and resource sharing. The purpose of the virtual machine is to configure the number of resources that are better suited to its real needs. Then, in order to deal with the dynamic change of application load in cloud environment, an optimized multi-tier application dynamic resource adjustment scheme is proposed, which mainly studies when resource adjustment is needed and how to adjust it. Finally, integrating the first two steps of the research, the paper puts forward the overall system architecture and module partition of cloud environment resource management runtime optimization, and uses the industry mainstream benchmark application RUBiS to test, The correctness and validity of the scheme are verified.
【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】