Combining computational and analytic model descriptions for evaluating embedded real-time systems

Industrial embedded real-time systems such as cabin communication systems (CCS) of modern (passenger) aircrafts [20], are highly complex. Verification of such complex systems is either hampered by the state space explosion problem, –attached to state-based verification,– or lacks a high-degree of detail, –due to the usage of some analytic method. For achieving a detailed analysis of system components, but still maintaining scalability of the procedure for computing overall properties of a system, our current research tackles the combination of executable and analytic model descriptions for the joint analysis of embedded real-time systems. As first step this report focuses on a combination of Timed Automata [2] with the MPA-method [8], where the latter is a relatively new compositional, analytical performance evaluation method based on the well-known Network Calculus [7]. As major outcome we present an efficient approach for converting analytic (sub)model descriptions of the MPA-method into networks of interacting Timed Automata.