Design And Implement Of B/S-based Automatic Deployment System For Openstack Cloud Platform

Since the AWS service of amazon achieved a huge success,companies based on the cloud platform facilities providing cloud services are appearing one after another.By building a cloud computing platform on the traditional IT infrastructure,IT resources or services can be sold online like ordinary goods and billed by the quantity.For traditional IDC suppliers,this improves the utilization rate of various resources and energy efficiency ratio.For users,cloud computing platforms reduce their operating costs and make it easier for them to build their own business systems.Among many cloud computing solutions,OpenStack,an open source cloud computing platform,is undoubtedly the most eye-catching.Since its open source,it has attracted many business giants such as IBM,RedHat and HP to join.The open source community is expanding continuously,and the technical precipitation is becoming more and more profound.Nowadays,OpenStack has become a synonym for open source cloud computing platform,and many companies have built their own public or private cloud platforms based on OpenStack technology.However,it has never been easy to build a cloud platform,mainly because OpenStack cluster is too large,the dependencies between components are too complex,and the configuration parameters are too many to manage.Therefore,how to quickly and efficiently deploy the large-scale OpenStack cloud computing platform has become an urgent problem to be solved.Based on the full investigation of OpenStack architecture model and the principle of main components,this paper designs a B/S architecture based cloud platform automation deployment subsystem.The system mainly consists of five parts,namely login management module,user management module,resource discovery module,system installation module and component installation module.In addition,the system also implements health check,cluster expansion,configuration item management and other extension functions.All functions can be operated through the Web page.For the collection and collation of configuration parameters,metadata is used to describe configuration items,so as to realize the increase or decrease of front-end configuration items without modifying background code.Adopting the idea of asynchronous programming,the message queue and task queue are used to receive and process internal requests of the system to avoid long waiting time.In software deployment,the components are loosely coupled to make the deployment process more intelligent.Based on the current popular B/S structure,this system is designed and implemented by Python Flask framework and AngularJS framework.After unit test and functional test,it is deployed to the production environment,which greatly improves the work efficiency of the deployer.