Software reliability accelerated testing method based on test coverage

In order to increase the effectiveness and efficiency of software reliability testing, an accelerated testing method based on test coverage is presented in this paper. The amount of reliability test cases is reduced by a heuristic test-suite reduction algorithm based on static analysis of test coverage. The testing process is accelerated by the strategy of not executing redundant test cases and compensating their execution time. The corresponding software reliability accelerated testing (SRAT) process is proposed. Also a novel two-dimensional NHPP software reliability growth model (SRGM) depending on testing time and test coverage simultaneously is developed to describe software reliability growth phenomenon in the SRAT process. The estimation method of its parameters is discussed. Finally the verification work using a case study has also been done and is introduced. Some conventional SRGMs including GO, MO and Yamada_Ohba_Osaki models are used as comparisons. Results of the experiment demonstrate clearly that the accelerated method can accelerate the reliability testing process dramatically and the proposed model is able to conduct feasible and accurate reliability assessment. Their feasibility and effectiveness are verified.