Combining Model Learning and Data Analysis to Generate Models of Component-Based Systems

摘要

Finding bugs in systems without model is well-known to be challenging and costly. But, most of today's developers think that writing models is also a hard and error-prone task. In this context, this paper addresses the problem of learning a model, from a component-based system, which captures and separates the behaviours of components and encodes their synchronisations. We present a passive model learning method called COnfECt to infer such models from execution traces in which no information is provided to identify components. We describe the two main steps of COnfECt in this paper and show some preliminary experimentations on real systems.