云计算中基于多维空间拓扑映射的资源机会复用技术研究

项目名称： 云计算中基于多维空间拓扑映射的资源机会复用技术研究

项目编号： No.61202113

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 计算机科学学科

项目作者： 钱柱中

作者单位： 南京大学

项目金额： 23万元

中文摘要： "按需服务"要求云平台为用户提供数量适当的资源，而资源复用则是云平台整合用户需求的基本途径，两者的本质目标都是通过保持云端合理的资源利用率，满足需求并降低成本。本项目以提高云平台资源利用率为基本出发点，以建立高效资源复用机制为核心，拟以准确刻画用户需求出发，研究包括资源需求拓扑的资源模型和需求映射机制，确立资源的概率使用模式，由此研究冲突约束的资源机会复用机制。在此资源复用框架下，研究该机制的核心资源调度技术：从满足用户请求的资源预分配角度，研究基于多维空间拓扑映射的均衡资源配置技术，实现分时复用的可伸缩资源配置机制，以达到整体最小化资源占用；从运行态资源调整角度，研究支持快速伸缩的自适应资源重配置技术，以适应负载波动并能保持系统高效率运行。上述成果将运用于我们现有云平台原型，并验证其正确性和有效性。该系统化的资源机会复用机制研究将促进资源利用率的提升，为云服务用户与云平台降低成本。

中文关键词： 云计算；虚拟化；资源配置；机会复用；

英文摘要： On-demand service requires the cloud platform providing suitable resource, while resource sharing is the basic method to consolidate the users' requests. Their essential goals are both reducing the running cost of resources through maintaining reasonable resource utilization. This project is to build an efficient resource sharing mechanism based on improving the resource utilization of cloud platform. We will investigate how to accurately represent the users' requests through resource model and request mapping mechanism including the resource topology. And we will investigate the opportunistic resource utility mode and collision restricted resource sharing mechanism. To reach the optimal resource utilization, we will then investigate the core resource allocation algorithms for the opportunistic resource sharing. From the view of pre-allocation resources to meet the requests, we will investigate multi-dimension topology mapping based balanced resource allocation technology to support time division multiplexing resource sharing, in order to maximum the resource utilization. From the view of running time resource adjusting, we will investigate self-adaption resource re-allocation that supporting scaling quickly of resources, so that it can adapt the load variation and maintain the high efficiency of the system. The

英文关键词： Cloud Computing；Virtualization；Recourse Configuration；Opportunistic Sharing；