Multiple Automatic Application Deployment Mechanisms Based On Cloud Computing Platform For Concurrent Multi-users

The research of future network has always been the focus of computer network technology. The future network will provide users with a more advanced network environment, and offer better support of the upper layer applications and services. However, it is difficult for many researchers in the application or service research and experiment to build a service validation experiment environments focused on Future Network due to the limit of their own resources or tools, and they can only build a low-performance local network environment. And researchers need to manually deploy a variety of software when they build a local environment, which is a time-consuming and labor-intensive process, and will seriously affect the efficiency of scientific research. This situation requires a service innovation experiment environment that can centrally manage resources and deploy tools, and an automatic deployment mechanism based on service innovation platform. However, with the increasing number of users of the service innovation cloud platform, the performance of the deployment mechanism will face more and more challenges. The user's multiple concurrent requests will lead to the increment of the response delay, the degradation of the system performance or even the collapse of the system. In order to solve these problems, this thesis will give three designs of the system architecture for concurrent multi-users, and through single mechanism, dynamic load balance and distributed systems to put forward the three implementions of deployment mechanism that can provide functions of the application deployment and the version change.The main contents of this thesis are as follows:1.Designing and realizing an automatic deployment mechanism with single server based on cloud computing platform for concurrent multi-users.This mechanism combines the Apache and the Passenger to change the system architecture, realizes the function of the application deployment when experimental environment is built and realizes the function of application redeployment for requirements change.2.Designing and realizing an automatic deployment mechanism with dynamic load balance based on cloud computing platform for concurrent multi-users. This mechanism combines the Nginx and a cluster of Puppets to change the system architecture, monitors the performance of servers to change the load balance strategy, realizes the function of the application deployment when experimental environment is built and realizes the function of application redeployment for requirements change.3 .Designing and realizing an automatic deployment mechanism with distributed systems based on cloud computing platform for concurrent multi-users. This mechanism combines the Git and Puppet to change the system architecture, realizes the automatic push of server codes, realizes the automatic gain of client codes, realizes the function of the application deployment when experimental environment is built and realizes the function of application redeployment for requirements change.In order to validate the availability of the system the correctness of the system under each mechanism, functional test cases and performance test cases are designed. The functional tests show that each mechanism can achieve the expected functional requirements. Performance testing shows the three deployment mechanisms can provide excellent performance when the number of concurrent requests is few. However, with the increasing number of concurrent requests, the performance of the mechanism with a single server has decreased quickly, the performance of mechanism with the dynamic load balance has decreased slightly and mechanism with distributed systems has always had an excellent performance.