Optimization And Research On Core Technology Of Software Testing Based On Cloud Computing

Software testing is a link with the characteristic of taking much time and power during the process of software development. The automation testing is an effective means to improve the efficiency of software testing while the effect is still relatively limited. With the development of cloud computing, it offers possibility for constructing virtual and reliable environment and realize the software testing which is convenient and low cost according to the characteristics of super large-scale, virtualization, high availability, universal and extension.This thesis research the key issues that to improving the capability of parallel test based on cloud computing. It focuses on the research of parallel test method. Through t he definition of test scenarios, test scene flow graph and parallel testing path to describe the relationship within the software testing function point. Through a semantic description of the test cases and use the semantic description of test case to generate test scenarios flow chart and parallel test path. Due to writing and maintaining test cases always cost too much, this thesis proposed to separate the test scripts and test data and configure test case dynamically.This thesis analyzes the existing algorithms for scheduling the resources of cloud testing. It finds that the existing scheduling algorithms has been relatively complete while most of them had not considered the characteristics of cloud testing. The dependence during test tasks and the precedence of test tasks are not considered. According to this, this thesis proposed a improved ant colony algorithm based on the characteristic of the cloud testing.Finally, it designs a experiment on the C loudStack platform for the proposed parallel test method and the method of generating test cases dynamically. This simulation proved that the proposed methods is efficient. It also design a experiment for the improved scheduling strategy on CloudSim platform. This simulation verified the algorithm combined with cloud testing characteristics and ant colony optimization performance well on execution time and load balancing.