基于OpenStack的业务云平台的动态可伸缩性研究与实现

发布时间：2018-05-04 06:57

本文选题：云计算 + OpenStack　；参考：《北京邮电大学》2014年硕士论文

【摘要】：目前,企业计算领域正在出现一种新的研究热点,就是云计算。这种计算模式允许用户根据自己的需要获取计算资源。从而当大量的请求突然到来时,企业能够良好地处理并且做到按需分配。倘若一个企业按照需求的最大值分配资源来适应随变化的时间而大幅度变动的计算服务请求,那么它必须做好巨额投入的预算。倘若它为了节省开销而走向另一个极端,即只分配资源的最小值,那么在突发性请求来临时,该企业可能面临难以预测的可怕后果。另一方面,云计算的发展产生了很多流行的开源云计算技术,OpenStack是典型的一种。OpenStack拥有商业云所具有的一切特点,同时它还包含部署简单,易于开发和扩展的优点,特别适合于企业内部和科研机构搭建私有云。因此,本文基于OpenStack开源云技术,针对现有的一个面向移动用户的多媒体服务平台,进行了业务云平台的动态可伸缩性研究,提出了一个分层次的云平台模型,并实现了相应的云扩展算法。本文首先介绍了云计算技术当前的研究情况,然后对虚拟化技术、它与云计算的关系和云计算的分类等进行了阐述,还对几种开源的IaaS云平台进行了比较。本文接下来简要地介绍了已有的多媒体服务平台,将它的服务进行拆分并做了简要的解析。本文的重点是OpenStack云的调研,目标是搭建一个多节点的OpenStack业务云平台,并基于它完成了具有高可伸缩性的云化业务执行平台的设计与实现。实验结果显示,这个业务云平台能够实现在时间限制和最小化能耗的条件下虚拟机资源的自动扩展机制,以保证高突发高波动性的负载请求下的服务质量并节约开销。
[Abstract]:At present, the enterprise computing field is emerging a new research hotspot, namely cloud computing. This computing model allows users to access computing resources according to their needs. Thus, when a large number of requests suddenly arrive, the enterprise can handle well and distribute according to demand. If an enterprise allocates the resources according to the maximum demand to adapt to the computing service request which changes greatly with the change of time, then it must make the budget of the huge investment. If it goes to the other extreme to save money, that is, to allocate only the minimum value of resources, the enterprise may face unpredictable and terrible consequences when sudden requests come. On the other hand, the development of cloud computing has produced many popular open source cloud computing technologies OpenStack is a typical. OpenStack has all the characteristics of the business cloud, but it also has the advantages of simple deployment, easy to develop and extend. Especially suitable for enterprises and scientific research institutions to build private clouds. Therefore, based on the OpenStack open source cloud technology, this paper studies the dynamic scalability of the service cloud platform for a mobile user oriented multimedia service platform, and proposes a hierarchical cloud platform model. The corresponding cloud expansion algorithm is implemented. This paper first introduces the current research situation of cloud computing technology, then describes the virtualization technology, the relationship between virtualization and cloud computing and the classification of cloud computing, and also compares several open source IaaS cloud platforms. Next, this paper briefly introduces the existing multimedia service platform, splits its services and makes a brief analysis. The focus of this paper is the research of OpenStack cloud, the goal is to build a multi-node OpenStack service cloud platform, and based on it, the design and implementation of cloud service execution platform with high scalability are completed. The experimental results show that the service cloud platform can automatically expand virtual machine resources under the condition of time limit and minimum energy consumption to ensure the quality of service under high burst and high volatility of load request and save the 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】：2014
【分类号】：TP393.07

【参考文献】