Using Coverage and Reachability Testing to Improve Concurrent Program Testing Quality

The testing of concurrent software is a challenging task. A number of different research approaches have investigated adaptation of the techniques and the criteria defined for sequential programs. A major problem with the testing of concurrent software that persists is the high application cost due to the large number of the synchronizations that are required and that must be executed during testing. In this paper we propose a complementary approach, using reachability testing, to guide the selection of the tests of all synchronization events according to a specific coverage criterion. The key concept is to take advantage of both coverage criteria, which are used to select test cases and also to guide the execution of new synchronizations, and reachability testing, which is used to select suitable synchronization events to be executed. An experimental study has been conducted and the results indicate that it is always advantageous to use this combined approach for the testing of concurrent software.