Optimized Test Data Generation for RESTful Web Service

To validate whether RESTful Web service systems could provide available services, it is indispensable to perform effective functional testing before large-scale deployment, where the pivotal and first step is to automatically generate valid test data. In this paper, we propose an optimized test data generation method for RESTful Web service systems. Mutation testing and genetic algorithm are utilized together to generate feasible and effective test data based on data type constraint extensions in the WADL specification. Then, Hadoop technology is well used to implement parallelized test data generation to promote efficiency of large-scale test data generation. Practical testing experiment results show advantages of our test data generation method. First, the number of feasible mutant candidates set and final test data set are effectively reduced. Second, Hadoop based parallelized test data generation has better efficiency and applicability for large-scale testing of RESTful Web service applications.