项目名称: 基于包簇映射的云数据中心资源管理模型构建及其若干关键技术研究
项目编号: No.61472256
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 自动化技术、计算机技术
项目作者: 陈世平
作者单位: 上海理工大学
项目金额: 82万元
中文摘要: 云计算取得了巨大成功,为服务于快速增长的用户群体,数据中心不断建立,但数据中心管理却面对扩展性、灵活性、综合性等多重挑战。本研究探讨数据中心管理技术,它将传统的以虚拟机为中心的管理模式转换为以包为中心的模式,其层次化的包-簇抽象模型通过递归定义方式将原本规模巨大的云基础资源分配问题分解为容易处理的小问题,以提高数据中心管理的扩展性;同时,通过所抽象的包和虚拟机的本地化资源分配模型,提高云基础资源的共享程度和资源利用率,进而提高数据中心资源管理的灵活性;此外,研究基于多重约束的资源分配问题解决方案的优化及多项相关关键技术以利于提高综合管理的能力和效率。研究的创新成果主要有:(1)基于包-簇映射的数据中心管理基础架构和工作机制;(2)能够处理多种实际需求和约束的综合管理框架;(3)一个建立于实验床基础上的基于包-簇概念的实验原型系统。
中文关键词: 云计算;数据中心管理;虚拟机;扩展性;灵活性
英文摘要: The enormous, ongoing success of cloud computing has promoted the establishment of large data centers to serve a rapidly expanding user base. However, resource management in data centers is facing grave multi-faceted challenges for scalability, flexibility, and comprehensiveness. This research project investigates new data center management technologies and proposes transformative shift from the conventional virtual-machine-centric management models to a new abstract model. Its layered clustering abstraction model divides the exponentially-complex resource allocation problem in data centers into a collection of much smaller problems in a recursive fashion based on the intrinsic structures in the client demand and the data-center networks, allowing flexible resource sharing through VM co-location. In addition, the proposed research addresses fundamental technologies for multi-constrained cloud resource optimization in order to improve manageability of data centers and their resource efficiency. The innovation of this project includes (1) the basic framework and operational mechanisms of large-scale data center management, (2) the comprehensive cloud management framework that can handle various realistic user demands and resource constraints, and (3) an experimental prototype.
英文关键词: Cloud Computing;Datacenter Management;Virtual Machine;Scalability;Flexibility