A model-based view onto testing: Criteria for the derivation of entry tests for integration testing

In many application domains, organization, cost, and risk considerations continue to lead to increasingly distributed system and software development processes. In these contexts, suppliers provide components, or entire subsystems, that are assembled by system integrators. One prominent, and tangible, example for such a development paradigm is the automotive domain where multiple embedded systems are integrated into a car (see Pretschner et al. 2007, Reiter 2010). For reasons of economy, suppliers aim at selling their subsystems to as many car manufacturers (usually and somewhat counterintuitively called original equipment manufacturers, or OEMs, in this context) as possible. This requires that their components work correctly in a multitude of different environments, which motivates thorough testing (and specification) of the components of a car under development. Each OEM, on the other hand, wants to make sure that the external components work as expected in its particular cars.