Mining behavioral specifications of distributed systems

Abstract

Software specifications provide explicit and high-level descriptions of a program ensuring a clear and consistent understanding of expected behavior. The absence of good specifications in real life software engineering processes have motivated research into automated techniques to 'mine' specifications from traces collected during actual executions of a software system. Current specification mining methods are largely limited to the analysis of sequential interactions between software components. This dissertation presents problems and methodologies in an attempt to advance the application of specification mining in two directions. First, it proposes methodologies and algorithms for mining specifications that account for concurrency and asynchronicity of processes in a distributed system. These methods are then coupled with a process class abstraction technique to produce simpler and more accurate specifications. The second advancement is an approach to mine difference specifications that elicit changes across versions and help to automatically adapt and maintain specifications of evolving software.