云计算资源管理中QoS保障机制研究

发布时间：2018-10-29 17:36

【摘要】：摘要：随着商业应用和科研领域的计算需求日益增长,云计算为满足这些计算需求提供了一种符合成本效益的的计算资源供给方式,并在工业界和学术界得到迅速发展和广泛应用。云数据中心将计算资源、存储资源和网络资源构建成动态的虚拟资源池,实现云资源的自动部署、动态扩展和按需分配,用户通过网络随时、随地、按需获得资源服务。然而,应用于商业环境的云计算需要面对复杂多样的用户需求,提供有质量保障的服务是云计算获得市场认可和商业推广的关键,因此,如何通过有效地用户任务分配和虚拟资源调度来保障用户QoS (Quality of Service,服务质量)是云计算资源管理中的重要问题。本文结合云计算的商业模式和技术特点,对云计算资源管理中的QoS保障机制进行研究并提出解决方案。首先通过讨论云计算QoS保障的关键问题,研究可量化的云计算QoS指标体系,并设计了提供QoS保障的云计算资源管理系统,其中包括任务在虚拟机资源上的分配和虚拟机资源在宿主物理主机上的调度。然后分别对系统中的两级调度问题分别深入研究：一方面根据用户服务请求的多样性,从优化成本的角度提出基于分级选址模型的资源分配优化模型,另一方面根据数据中心中物理资源复杂异构的特点,从提高资源利用率的角度提出基于背包问题的虚拟资源动态调度模型,并给出改进蚁群算法的资源调度策略。本文通过将用户的QoS需求作为约束条件应用于资源管理模型的建模和设计中,形成了最大化经济效益的云计算资源管理QoS保障机制。最后,通过扩展CloudSim云计算仿真平台,本文对云计算资源管理的模型及算法进行了实验验证和对比,证明了本文提出的算法的可行性和有效性,实现了在保障QoS的同时最优化运营成本和资源利用率,为云计算资源管理中的QoS保障研究提供了有效的借鉴方案。
[Abstract]:Absrtact: with the increasing demand for computing in commercial applications and scientific research, cloud computing provides a cost-effective way to supply computing resources to meet these computing needs. And in industry and academia has been rapid development and wide application. Cloud data center will construct dynamic virtual resource pool of computing resources, storage resources and network resources, realize the automatic deployment of cloud resources, dynamic expansion and distribution according to demand, and users can obtain resources services at any time, anywhere and on demand through the network. However, the application of cloud computing in the business environment needs to face complex and diverse user needs. Providing quality assurance services is the key for cloud computing to gain market recognition and business promotion. It is an important problem in cloud computing resource management that how to ensure the quality of service of user QoS (Quality of Service, through effective user task allocation and virtual resource scheduling. Combined with the business model and technical characteristics of cloud computing, this paper studies the QoS guarantee mechanism in cloud computing resource management and puts forward solutions. Firstly, through discussing the key issues of cloud computing QoS guarantee, the quantifiable index system of cloud computing QoS is studied, and the cloud computing resource management system which provides QoS guarantee is designed. It includes task allocation on virtual machine resource and scheduling of virtual machine resource on host physical host. Then the two-level scheduling problem in the system is studied in depth. On the one hand, according to the diversity of user service requests, a resource allocation optimization model based on hierarchical location model is proposed from the point of view of optimizing cost. On the other hand, according to the complexity and heterogeneity of physical resources in data centers, a dynamic scheduling model of virtual resources based on knapsack problem is proposed from the point of view of improving resource utilization, and a resource scheduling strategy based on improved ant colony algorithm is presented. In this paper, the QoS requirements of users are applied to the modeling and design of resource management model, and the QoS guarantee mechanism of cloud computing resource management is formed to maximize the economic benefits. Finally, by extending the CloudSim cloud computing simulation platform, this paper verifies and compares the model and algorithm of cloud computing resource management, and proves the feasibility and effectiveness of the proposed algorithm. It realizes the optimization of operating cost and resource utilization while guaranteeing QoS, and provides an effective reference scheme for the research of QoS guarantee in cloud computing resource management.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09

【参考文献】