A Formal Approach to Enforcing Consistency in Self-adaptive Systems

The ability of systems to adapt is increasingly seen as a necessary underlying capability for modern software systems. The resulting self-adaptive systems are not only supposed to cope with changes, but must also preserve their consistency. To deal with such challenges in a systematic way, the design of self-adaptive systems needs to be put on a formal basis. In this paper, we argue for the benefits of a formal yet extensible approach to behavioural adaptations of component-based system architectures. This approach provides the usage of alternative adaptation processes rather than being limited to a single one. The application of Coloured Petri Nets for modelling and analysing the adaptation processes proves to be useful to trust consistency preservation.