Jun Li

Testing of Distributed and Component-Based Systems

The cost and complexity of developing and maintaining test suites to achieve defect-free applications often exceeds that of developing and maintaining these applications. It becomes much more significant for large-scale distributed applications that keep evolving within their life span. Our research effort is to build a test automation infrastructure and its tool support to reduce the cost and complexity of developing and maintaining such complex test suites by automating: test cases generation, test effectiveness evaluation, regression testing, and impact change determination.

In particular to distributed and component-based applications, we adopt the two-folded approach: first, to exploit the distributed systems monitoring techniques that we have developed to perform test execution runtime monitoring, and to collect and store test execution information; and secondly, to augment the IDL for test design specifications and automate the generation of test cases and the test harness. Overall, our approach is to explore the linkages among three parties altogether: the application under test, its behavior model, and the underlying middleware runtime infrastructure.