云环境中虚拟机性能互扰度量和预测方法及其应用研究

发布时间：2018-05-03 07:40

本文选题：云计算 + 资源分配　；参考：《东北大学》2013年硕士论文

【摘要】：云计算作为一种新的服务交付和使用模式,旨在为用户提供安全、快速、便捷的数据存储与网络计算服务。同时,动态扩展和按需付费等特性使得它很好地满足了控制服务成本和降低能耗的需求,因而具有广阔的市场前景。在云计算数据中心,资源管理和分配的最小粒度是虚拟机,通常用户都希望其能够稳定地运行,因此如何有效地实现虚拟机分配成为研究热点。然而在目前的研究工作大都忽略了虚拟机性能互扰对于虚拟机承载应用的性能影响,无法保证虚拟机分配结果的有效性。为了保证云计算中资源分配的有效性,本文在研究云环境中资源分配和虚拟机间性能互扰度量等相关工作的基础上,针对如何客观有效度量预测虚拟机间性能互扰度并将其应用于资源分配的问题,设计了一个云环境中虚拟机性能互扰度量和预测方法。该方法包含三个步骤,第一步是构建虚拟机互扰度模型并进行模型参数求解,第二步是利用层次聚类算法对虚拟机负载模式进行挖掘,第三步是将当前虚拟机负载模式与虚拟机负载模式聚类结果进行匹配,利用与匹配结果相应的性能互扰度模型完成当前虚拟机性能互扰度预测并对分配方案进行互扰度评价。本文最后通过一系列虚拟机部署实验验证所提出的虚拟机性能互扰度量和预测方法在保证服务质量方面的有效性。
[Abstract]:As a new mode of service delivery and use, cloud computing is designed to provide users with secure, fast and convenient data storage and network computing services. At the same time, the characteristics of dynamic expansion and demand payment make it well meet the cost of controlling service and reduce the demand for energy consumption, so it has a broad market prospect.
In the cloud computing data center, the minimum granularity of resource management and allocation is virtual machine, and usually users want it to run steadily. Therefore, how to effectively implement virtual machine allocation has become a hot topic. However, most of the current research work ignores the performance impact of VM performance interference to the virtual machine application. Verify the effectiveness of the virtual machine allocation results.
In order to ensure the effectiveness of resource allocation in cloud computing, this paper designs a virtual machine performance in a cloud environment on the basis of research on resource allocation and performance cross disturbance measurement between virtual machines in the cloud environment and how to objectively and effectively measure the performance interinterference between virtual machines and apply it to resource allocation. The method consists of three steps. The first step is to build a virtual machine mutual disturbance model and solve the model parameters. The second step is to use the hierarchical clustering algorithm to mine the virtual machine load mode, and the third step is to match the current virtual machine load mode with the virtual machine load pattern clustering results. With the result of the corresponding performance interference degree model, the performance interference degree of the virtual machine is predicted and the interference degree of the allocation scheme is evaluated. Finally, a series of virtual machine deployment experiments are carried out to verify the effectiveness of the proposed virtual machine performance interference measurement and prediction method to ensure the quality of service.

【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP302

【参考文献】