| As an important means of software quality assurance, software testing is playingan increasingly important role. Traditional software testing improves test efficiencyby optimizing test cases, but this will reduce the discovery of defects and affect thetest quality. Parallel testing technology was introduced to solve this problem.However, parallel testing needs to build a clustered environment, which costs a lot.The emergence of cloud computing provides new possibilities for automatedtesting. The low cost and strong scalability of cloud computing provides a new wayfor building a cost-effective automated testing environment, but also presents newchallenges. In a cloud computing environment, the traditional test cases requireparallel processing, while the demand of fully testing requires automated generationof a large number of test data.This paper focuses on a study of parallel test case generation based on cloudcomputing, which mainly includes the parallel processing of test scripts, and anautomated test data generation method. Firstly, according to scene flow graph，parallel test paths are generated automatically by depth-first traversal algorithm，parallel test path scripts and initial test data can be obtained at the same time. Theneffective test data, which can traverse target paths and replace the initial test data inscripts, are generated automatically using Search Based Software Testing (SBST)method.Based on open-source automated web testing tool Selenium, cloud computingframework CloudStack, and JAVA web development framework SSH, this paperbuild an automated cloud testing prototype system. It has functions including testingprojects management, test scripts submitting, test data generation, scheduling test task,and collecting test results.Finally, through some experiments, this paper verify the effectiveness of theproposed method, and proves that the automated cloud testing can improve theefficiency of software testing and save the cost of software testing. |
PDF Full Text Request