Research And Implementation Of Universal Software Reliability Cloud Test Platform

With the development of software technology, a variety of software and systems become more and more complex, and populace increasingly demand high quality software. Software testing is one of the most important means to ensure software quality. However, many of the existing test software are stand-alone, and can't adapt to a variety of test scenarios any more due to the software system has become more complex and diverse. And the development of software technology in recent years brings to a lot of new technologies, which not only changed the pattern of the software industry, but also brought new ideas for the development of test software, such as Cloud Computing technology.Cloud Computing has a lot of advantages such as low cost, pay-per-use, security and reliability. If the existing stand-alone test software can run on the cloud, test software can achive the various advantages of the Cloud Computing, such as increasing the scalability of test software,reducing the cost of testing, etc. Therefore, how to implement a cloud platform for test software is the main research content. Besides, function and performance of the platfrorm are verified by software reliability testing.Some of the existing stand-alone test software are evaluated, which are deficient in extensibility, reliability, test cost and flexibility, etc. In order to propose a distributed cloud test platform which can overcome these shortcomings and be used for a variety of test software simultaneously.This theis presents a cloud test platform solution based on the YARN resource management framework, including storage of test results, real-time processing of test results, test task submission, and Web module. A cloud test platform prototype system is developed through a large number of research, which uses open-source and self-developed components. In addtion, JMeter, a stand-alone test software, is run on the cloud test platform, and functional verification and performance experiment are performd. The experiment result shows that the test software has the ability of linear scalability on the cloud platform, and the performance and stability of the test task are better than the original running mode.