View article

Variability is essential for adaptive software systems, because it captures the space of alternative adaptations a system is capable of when it needs to adapt. In this work, we propose to capture variability for an adaptation space in terms of a three dimensional model. The first dimension captures requirements through goals and reflects all possible ways of achieving these goals. The second dimension captures supported variations of a system’s architectural structure, modeled in terms of connectors and components. The third dimension describes supported system behaviors, by modeling possible sequences for goal fulfillment and task execution. Of course, the three dimensions of a variability model are inter-twined as choices made with respect to one dimension have impact on the other two. Therefore, we propose an incremental design methodology for variability models that keeps the three dimensions …