Simplified Validation of Emergent Systems through Automata Learning-Based Testing

摘要

In this paper we present a novel approach to the test-based validation of complex heterogeneous applications which is tailored to simplify the requirements for the responsible personnel. Key to our approach is to automate the corresponding testing procedure by means of active automata learning. This replaces typical prerequisites like manual test construction or the provision of adequate test models by the definition of an adequate learning alphabet. In practice this typically means by providing representative data for the parameters of the relevant API calls. Besides its simplicity this approach also guarantees that the testing procedure automatically adapts to system modifications, which makes it an ideal tool for dealing with evolving systems of unknown emergent behaviour, as will be illlustrated along the Online Conference System (OCS), a model-driven and service-oriented online conference manuscript submission and review system.