As test suites grow larger and more complex, the time and resources required to execute the test suite also grows accordingly. This is particularly true for companies developing diverse product lines, who must execute portions of the test suite on compatible hardware units with the features required by those test cases. We have recently begun to explore the topic of how to efficiently parallelize the execution of test cases across a pool of available hardware units. We propose a search-based test scheduling method that attempts to optimize the estimated execution time and utilization of the hardware pool. In this talk, we will present our initial ideas, the results from an early pilot study, and our plans to integrate this approach into the test execution infrastructure at a Software Center partner company.
Speaker: Gregory Gay