Model-based integration and testing in practice
N.C.W.M. Braspenning, J.M. van de Mortel Fronczak, H.A.J. Neerhof, J.E. Rooda · TU/e Research Portal · 2007
High-tech multidisciplinary systems like wafer scanners, electronic microscopes and high-speed printers are becoming more complex every day. Growing system complexity also increases the effort (in terms of lead time, cost, resources) needed for the, so-called, integration and test phases. During these phases, the system is integrated by combining component realizations and, subsequently, tested against the system requirements. Existing industrial practice shows that the main effort of system development is shifting from the design and implementation phases to the integration and test phases [32], in which finding and fixing problems can be up to 100 times more expensive than in the earlier requirements and design phases [14]. As a result, the negative influence of the integration and test phases on the Time–Quality–Cost (T-Q-C) business drivers of ASML (see Chapter 3) is continuously growing and this trend should be countered. Literature reports a wealth of research proposing a model-based way of working to counter the increase of system development effort, like requirements modeling [34], model-based design [78], model-based code generation [60], hardware-software cosimulation [108], and model-based testing [33]; see also Chapters 9 and 10. In most cases, however, these model-based techniques are investigated in isolation, and little work is reported on combining these techniques into an overall method. Although model-based systems engineering [89] and OMG’s model-driven architecture [72] (for software only systems) are such overall model-based methods, these methods mainly focus on the requirements, design, and implementation phases, rather than on the integration and test phases. Furthermore, literature barely mentions realistic industrial