Research And Application On Performance Testing Technologies Of Distributed System Based On Cloud Computing

With the rapid development of computer and network technologies, software systemgradually shift from centralized structure to distributed structure. Software testing is the mainmeans to evaluate the software quality, and performance testing is a key part of software testing.For the traditional test platform, there are some problems for distribute system test such as moreredundant test, more difficulty of locating the performance bottleneck. Cloud computing providesa new way for performance testing of distributed system, which has the characteristics ofon-demand service, generalization access, resource pooling, flexible service, billing services.The background of this paper is the project from Shanghai Municipal Education Commissionnamed Research on key technologies of software parallel test based on cloud computing, whichfocuses on the automated testing technologies of the software performance in cloud environment.In the process of performance testing of the distributed system, the testing number withexponential growth usually leads to the insufficiency of testing resources, so a performance testmethod based on decision tree in cloud environment is proposed，which is based on decision treemodel, use the testing resources of cloud environment to describe the generating of test cases, testscripts running and locating performance bottlenecks in the performance test process.On the other hand, to resolve the deficiency of scalability of the performance testingenvironment in the traditional test platform, a cloud-based distributed system performance testingplatform prototype system is built in this paper. It realized automated configuration and dynamicexpansion of performance testing environment in the cloud environment, which based onCloudStack, an open source cloud computing framework, LoadRunner, an automated performancetesting tool, SSH, a lightweight web development framework, JavaScript library and JQuerytechnology. The testing platform mainly consist of the modules of submission of test tasks, imagesmatching of virtual machine, automated configuration of test environment, cloud resourcescheduling, collecting the test results. Finally, an experiment is performed to verify the feasibility and effectiveness of the proposedperformance test method, which improved the efficiency of performance testing.