方舱计算  

Cabin computing

作　　者：蒋昌俊[1,2] 丁志军[1,2] 喻剑[1,2] 章昭辉[1,2] 闫春钢[1,2] 张亚英[1,2] 王鹏伟[1,2] Changjun JIANG;Zhijun DING;Jian YU;Zhaohui ZHANG;Chungang YAN;Yaying ZHANG;Pengwei WANG(Key Laboratory of Embedded System and Service Computing,Ministry of Education,Tongji University,Shanghai 201804,China;Network Information Service Shanghai Engineering Research Center,Tongji University,Shanghai 201804,China)

机构地区：[1]同济大学嵌入式系统与服务计算教育部重点实验室,上海201804 [2]同济大学网络信息服务上海市工程研究中心,上海201804

出　　处：《中国科学：信息科学》2021年第8期1233-1254,共22页Scientia Sinica(Informationis)

基　　金：国家重点研发计划(批准号:2018YFB2100800)资助项目。

摘　　要：随着信息技术创新日新月异,数字化、网络化、智能化深入发展,新应用层出不穷、新业态蓬勃发展,对业务系统的敏捷构造和持续运维提出了更高的要求.为此,本文提出了一种新的计算模式:方舱计算.所谓方舱计算,是通过网络访问的,面向IT任务全生命周期的跨域资源配置和协同的计算集成环境.其核心是"方舱专用机动、资源跨域伸缩、系统运维自治".本文给出了方舱计算系统结构及其工作原理,方舱计算系统主要由方舱生成与管理系统、跨域资源管理系统、虚拟数据中心系统和若干网关(方舱网关、虚拟数据中心网关、跨域资源网关等)组成.进而,给出了适于方舱计算的资源分配最优化问题定义,在兼顾数据资源、计算资源和存储资源等的基础上,实现最小化资源成本.通过规划求解器与近似优化算法仿真实验,表明本文的方法能够兼顾数据资源和物理资源的分布,实现方舱资源的优化配置.With the rapid development of information technology innovation, digitalization, networking and intelligence are deeply developing. New applications are emerging and new industries are flourishing, which puts forward higher requirements for the agile construction and continuous operation and maintenance of business systems. This paper proposes a new computing model called cabin computing. Cabin computing is the crossdomain resource configuration and collaborative computing integration environment for the full life cycle of IT tasks accessed through the Internet. Its core idea is "mobile special cabins, flexible cross-domain resource, and autonomous system operation and maintenance". This paper presents the architecture and working principle of the cabin computing system. It is mainly composed of cabin generation and management system, cross-domain resource management system, virtual data center system and several gateways(cabin gateway, virtual data center gateway, cross-domain resource gateway, etc.). Furthermore, the definition of resource allocation optimization in cabin computing is stated. Taking data resources, computing resources and storage resources into account, the resource cost is minimized. The simulation experiments of the programming solver and approximate optimization algorithm demonstrate that the proposed approach can realize the optimal allocation of cabin resources considering the distribution of both data resources and physical resources.

关 键 词：方舱 虚拟数据中心 跨域资源管理 资源分配 资源分布图 资源目录 

分 类 号：TP393.09[自动化与计算机技术—计算机应用技术]