云环境下商业驱动的能力规划模型的研究与实现

发布时间：2018-03-13 03:02

本文选题：云计算　切入点：能力管理　出处：《北京邮电大学》2013年硕士论文　论文类型：学位论文

【摘要】：云计算采用创新的计算模式,通过互联网为客户提供强大的计算能力和多样的信息服务。云计算的商业模式使计算和服务可以取用自由、按量付费。云计算融合了虚拟化、自动化和标准化等技术和特性。云计算供应商借助虚拟化技术的伸缩性和灵活性,提高了资源利用率,简化了服务和资源的管理和维护；并使用自动化技术,将资源封装交付给客户,实现环境的自动部署,减少了数据中心的运营成本；通过标准化,方便了服务的交付和开发,便于应用的复用。云计算的出现大大地改变了IT资源的消耗模型。通过使用云计算服务,客户可以根据需求灵活地增加或减少从云平台处获得的计算能力,并且只需要支付使用了的资源费用,无需再烦恼建设、运行和维护IT基础设施的成本问题。虽然云计算供应商大大减轻了企业的IT基础设施负担,但是如果要成本最小化,企业在购买云计算资源时还必须进行合理的能力规划。针对云计算上述特点和存在的问题,本文的工作主要分为如下几个方面： (1)以Amazon EC2为代表分析了目前云平台的三种主导购买方式——预约购买方式、按需购买方式和现货购买方式——下计算资源性能和价格的特点,其中着重分析了现货购买方式下Amazon EC2给出价格的规律特点。 (2)对云计算能力规划进行了需求分析,提出了长期能力规划和短期能力规划的设计理论。分别分析了长期能力规划和短期能力规划的作用和目标。 (3)设计了一个云计算长期能力规划数学模型。对模块的设计思想、解决的问题、实现的目标进行了详细的介绍。通过与其他已有的云计算能力规划模型进行性能比较和评估,总结得出本长期能力规划模型的特点和优势。 (4)设计了一个云计算短期能力规划数学模型,并以该数学模型为基础提出来一个云计算短期能力规划算法。并对该算法的基础思想分支限界法进行了详细的介绍。并详细介绍了设计的限界函数,并通过数学证明验证了其正确性。最后对该算法进行了编码实现。通过与其他算法进行性能比较和评估,总结出该短期能力规划在性能和时间上的特点和优势。 (5)提出了一个自动能力管理系统的架构。可以根据这个架构去实现一个完整的自动化能力管理系统。本文为购买云计算资源满足应用需求的客户提出了一套完整系统的能力规划方法。长期能力规划帮助客户在保证应用程序服务质量的基础上实现尽可能降低购买成本。短期能力规划帮助客户解决不可预测的应用程序需求问题,对长期能力规划起到了调整作用。实验表明,本文设计的两种能力规划模型在降低费用成本和时间成本上都有一定的优势。
[Abstract]:Cloud computing uses innovative computing models that provide customers with powerful computing power and diverse information services through the Internet. Cloud computing business models allow computing and services to be freely available and pay-as-you-go. Cloud computing incorporates virtualization. Technologies and features such as automation and standardization. Cloud computing vendors increase resource utilization, simplify the management and maintenance of services and resources, and use automation technologies by virtue of the scalability and flexibility of virtualization technologies, The resource encapsulation is delivered to the customer to realize the automatic deployment of the environment, and the operation cost of the data center is reduced. Through standardization, the service delivery and development are facilitated, and the reuse of the application is facilitated. The advent of cloud computing has dramatically changed the consumption model of IT resources. By using cloud computing services, customers can flexibly increase or reduce computing power from the cloud platform as needed, and pay only the cost of the resources used. There is no need to worry about the cost of building, running, and maintaining IT infrastructure. While cloud computing vendors greatly reduce the IT infrastructure burden on the enterprise, if the costs are to be minimized, Enterprises in the purchase of cloud computing resources must also be reasonable capacity planning. In view of the above characteristics and problems of cloud computing, the work of this paper is divided into the following aspects:. 1) taking Amazon EC2 as the representative, the paper analyzes the characteristics of the three dominant purchasing modes of cloud platform at present-booking purchasing, on-demand purchasing and spot purchasing-calculating the performance and price of resources. In this paper, the characteristics of the price given by Amazon EC2 under spot purchase are emphatically analyzed. 2) based on the requirement analysis of cloud computing capability planning, the design theory of long term capability planning and short term capability planning is put forward, and the roles and objectives of long term capability planning and short term capacity planning are analyzed respectively. (3) A mathematical model of cloud computing capability planning is designed. The design ideas, problems solved and objectives achieved are introduced in detail, and the performance is compared and evaluated with other existing cloud computing capability planning models. The characteristics and advantages of this long-term capability planning model are summarized. A short-term capability planning mathematical model for cloud computing is designed. Based on the mathematical model, a short-term capability planning algorithm for cloud computing is proposed. Finally, the algorithm is coded and implemented. By comparing and evaluating the performance with other algorithms, the characteristics and advantages of the short-term capability planning in performance and time are summarized. The architecture of an automated capability management system is presented, according to which a complete automatic capability management system can be implemented. This paper presents a complete and systematic capability planning method for customers who purchase cloud computing resources to meet their application requirements. Long-term capability planning helps customers to reduce purchases as much as possible on the basis of ensuring the quality of service of the application. Buy cost. Short-term capability planning helps customers solve unpredictable application requirements, The experiment shows that the two capacity planning models designed in this paper have some advantages in reducing cost and time cost.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F49;TP393.09

【相似文献】