The Design And Implementation Of Public Computing Management System Based On Docker

As a popular research field in academia and business,cloud computing is becoming known and valued by more and more people in recent years.According to the division of cloud computing service model,cloud computing can be divided into three levels of services:software as a service(SaaS),platform as a service(PaaS),and infrastructure as a service(IaaS).The cloud computing platform provides a large number of services,integrates loose functions,and manages,schedules,and monitors resources through a unified interface.Among them,based on the cloud service platform provided by the PaaS layer,application developers can develop,test,and deploy applications quickly and easily,operation and maintenance personnel can also easily perform operation and maintenance work,and the application is hosted on the PaaS layer.In the cloud computing platform,the purpose of shortening the development cycle and reducing the operation and maintenance costs can be achieved,and the work done by a large number of operation and maintenance personnel in platform construction and maintenance can be saved.With the rapid development of cloud computing,Docker,as an emerging lightweight container engine,provides reliable support for the construction of cloud platforms.Early cloud computing platform scheduling method using virtual machines to achieve,and now more use of container technology cloud platform to replace virtual machines,this article designed to use Docker as the underlying scheduling unit.The Docker technology,released in 2013,has been built out of the box on container virtualization.By packaging the runtime environment and applications together,it solves the environmental dependencies of the deployment.And it leverages the Linux Container(LXC)technology to implement application isolation.Traditional virtual machine technology implements hardware-level virtualization,while Docker implements operating system-level virtualization.So compared to virtual machines,Docker containers are lighter,have smaller granularity,and are more portable.For the PaaS platform,Docker can provide standardized packaging deployment and computing resource isolation capabilities,making the PaaS platform more simple and efficient.In this paper,a Docker-based public service computing management platform is designed based on Docker container technology.The system objectives and functional requirements were analyzed,the overall architecture of the system was designed,and various functional modules were implemented.The platform addresses the many limitations of traditional computing platform management methods,adopting the way of packaging applications in Docker containers,and deploying multiple applications on a cluster to facilitate operation and maintenance management.The platform provides multiple types of basic images for users to use,allowing users to upload application packages to build their own images.At the same time,the platform uses docker-compose to provide a template for application composition,which facilitates one-click construction.Using the mature Python Web application development framework Flask to write a Docker-based public service computing management platform.The scheduling of computational resource scheduling and containers is implemented using Docker's integrated Swarm tool.Swarm's goal is to efficiently share hardware resources among different application frameworks,and to simplify its own scheduling logic to make it as compatible and extensible as possible to ensure robustness in large-scale cluster use environments.And universal applicability to various possible computing frameworks.Using the Swarm resource manager to arbitrate different schedulers,we will enter a dynamic partitioning/elastic sharing model where all applications can use the node's public pool to safely and maximize resources.