Abstract. The design of critical embedded real-time systems requires high confidence in the architecture and the implemented functionalities. Classically, such functions are supported on a single monoprocessor, behavior of which is completely predictable, while the execution is totally deterministic. With the growing complexity of systems, it becomes quite unavoidable to implement these systems on distributed architectures. Scaling from mono to multiprocessor raises several issues which we address in this paper: we propose a simple executive model based on time triggered paradigm and an automated approach to allocate real-time tasks based on constraint resolution techniques. We illustrate the method on an industrial case study. Key words: Distributed systems, Constraint programming, Industrial case study