Research And Implementation Of Reliability Test Cases Generation Methods Based On HTTP

With the development of information technology, Web applications have been involved into all areas of our life, and the government,companies and consumers increasingly rely on the Internet. These changes make it more important to ensure the reliability of Web applications. Software testing is a critical stage in evaluating and improving the reliability of Web applications. Manual testing will cost a lot of human resource and time, but it produce little effect. So automated testing emerged. The generation of test cases is one of the most important parts in automated testing. Generation methods of Web applications reliability test cases are investigated in this paper.There are two methods respectively based on record or log analysis.It depends on whether the application has enough and effective access logs to choose one appropriate method. The research results of this thesis are as follows: 1) A developer tool of Chrome browser is implemented. It can record the HTTP or HTTPs requests sent by the Chrome browser. In addition, it has the functions of configuring record parameters,identifying dynamic parameters, editing scripts, downloading scripts and so on. 2) The Markov chain model for Web applications is defined. It involves the definition of state, transference relationship between states and probability of transference. 3) An algorithm to determine the adequacy of the test is proposed. This algorithm calculates a quantitative value of the difference between two Markov chains. This value can determine if the test is adequate or the test could be terminated.The experiment results show that both the Chrome extension based on record and the test cases generation module based on log analysis can satisfy their own function requirement and the determinant algorithm is more efficient than the Euclidean distance.