Modeling dependable systems: what can model driven development contribute and what likely not?

Dependability of complex systems is a critical system-level issue, which is still insufficiently tackled today. We propose the adoption of an aggressive model-driven development paradigm (AMDD) to adequately capture globally scoped issues like interoperability and compatibility. AMDD moves most of the recurring problems of compatibility and consistency of a system's design, implementation and evolution from the coding and integration level to the modeling level. In particular it enables an economical treatment of cross-system issues. This approach has proven useful already for the design, realization and test of complex distributed applications. In this paper we discuss the potential and benefits of AMDD and present examples where it is already adopted