云应用部署配置模型  

Deployment and Configuration Model of Cloud-Based Applications

作　　者：姬源 黄育松[1] 谢冬[1] 王向东[1] 

机构地区：[1]贵州电网公司电力调度控制中心,贵阳550002

出　　处：《计算机系统应用》2016年第5期13-18,共6页Computer Systems & Applications

摘　　要：云计算已经成为广泛使用的计算范型,越来越多的大规模分布式系统已经或正在向云平台部署和迁移.用户在部署和管理维护应用系统时通常需要管理底层基础设施资源细节,或者使用平台提供方的应用部署和管理服务,前者使得应用部署和运行时管理易于出错且费时费力,而后者则降低了系统管理的灵活性,很难满足用户的个性化需求.针对这一问题,本文提出了一种高层抽象模型来描述云应用的部署配置和管理需求.需求模型采用声明式机制定义期望的系统状态,而无需描述实现目标状态所需的执行步骤和细节.本文基于开源云计算平台Open Stack和自动化配置管理工具Puppet进行了原型实现,通过一个应用案例验证模型的有效性.Cloud computing has been a popular computing paradigm, and more and more large-scale distributed systems are migrating or have been migrated to cloud platforms. When deploying and managing IT systems in cloud environments, system owners have to manage the low-level details of the infrastructures, or leverage the services of platforms. Managing low-level details is time consuming and error-prone, and leveraging the services is less flexible and more difficult to tailor to user requirements. To address these issues, we propose a high-level abstraction model for cloud-based systems to specify their requirements. The requirement model describes the expected system states in a declarative way, without introducing the low-level steps and details about how to reach the states. We implemented a prototype based on open source cloud platform Open Stack and configuration management tool Puppet, and the effectiveness of the requirement model is evaluated through a case study.

关 键 词：云计算 需求模型 配置管理 部署 

分 类 号：TP393.09[自动化与计算机技术—计算机应用技术] TM76[自动化与计算机技术—计算机科学与技术]