云计算数据中心带宽成本优化算法研究

发布时间：2019-02-15 10:44

【摘要】：云计算技术近几年来得到长足的发展,无论是从硬件上的提升,还是计算性能的提升,都给互联网产业带来巨大的变化。全球云计算服务技术逐步发展,容器技术成为应用在云计算领域的主流技术之一。云计算技术中的虚拟化很好的实现将硬件资源统一管理并且给不同的用户提供不同的服务。虚拟化技术是对云计算数据中心所含有的硬件资源等基础设施资源实现共享和复用。数据中心随着时代的变化,并且伴随着大量新型需求的出现,云计算平台下的数据中心不得不在新的要求下做出相应的改变。在这其中,跨地域的数据中心的建设尤其显现的突出重要。尤其在商业环境下,如何能够在保证业务质量的情况下降低云计算平台下的数据中心的成本,成为一个尤为重要的课题。针对上述情况,优化数据中心的带宽成本问题成为一个很重要的研究热点,基于可延迟大数据窗口,在多延迟窗口下的大数据传输问题。提出多延迟窗口下带宽成本优化的数学模型。在此基础上研究了多延迟窗口下的在线算法,据此设计了 Multi-Heuristic Smoothing Algorithm(MHSA)并证明其竞争比是在2(1 — (1 - 1/D)D)到2之间。本文研究了多延迟窗口情况下的云平台在线算法的设计。在设计云平台算法时,本文提出成本函数Φ(Φ),采用设计Φ(Φ)选择最优可执行数据中心,最后达到优化成本的效果。基于Φ(Φ),设计在固定延迟下 Single-Dynamic Self-adaption Algorithm (SDSA)和多延迟下的Multi-Dynamic Self-adaption Algorithm (MDSA)并且最后模拟实验证明其优化的效果。
[Abstract]:Cloud computing technology has been greatly developed in recent years, both in terms of hardware and computing performance, all of which bring tremendous changes to the Internet industry. With the development of global cloud computing service technology, container technology has become one of the mainstream technologies in cloud computing. Virtualization in cloud computing technology is a good way to manage hardware resources and provide different services to different users. Virtualization technology is to share and reuse infrastructure resources such as hardware resources contained in cloud computing data centers. Data centers change with the times, and with the emergence of a large number of new requirements, the data centers in cloud computing platform have to make corresponding changes under the new requirements. Among them, the construction of cross-regional data center is particularly important. Especially in the business environment, how to reduce the cost of data center in cloud computing platform under the guarantee of business quality has become a particularly important issue. In view of the above, optimizing the bandwidth cost of the data center has become a very important research hotspot, which is based on the delay big data window, and then the big data transmission problem under the multi-delay window. A mathematical model of bandwidth cost optimization for multi-delay windows is proposed. Based on this, the online algorithm for multi-delay window is studied. Based on this, the Multi-Heuristic Smoothing Algorithm (MHSA) is designed and its competition ratio is proved to be between 2 (1- (-1 / 1) D) to 2. In this paper, the design of online algorithm for cloud platform with multiple delay windows is studied. In the design of cloud platform algorithm, the cost function 桅 (桅) is proposed, and the optimal executable data center is selected by designing 桅 (桅). Finally, the effect of optimizing cost is achieved. Based on 桅 (桅), the Single-Dynamic Self-adaption Algorithm (SDSA) with fixed delay and Multi-Dynamic Self-adaption Algorithm (MDSA) with multiple delays are designed, and the simulation results show that the optimization is effective.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP308;TP393.09

【参考文献】