Towards Translating FSM-SADF to Timed Automata

Dataflow formalisms play a significant role in the areas of design and analysis of embedded streaming applications. These formalisms can roughly be split into static and dynamic ones. Static dataflow formalisms are highly analyzable, but due to their static nature are not able to capture the dynamism inherent to modern embedded streaming applications. Dynamic dataflow formalisms on the other hand provide a sufficient level of expressiveness to capture the application dynamism at the cost of reduced analyzability. The recently introduced finite state machine-based scenario aware dataflow (FSM-SADF) formalism provides a good trade-off between expressiveness and analyzability. This paper reports on the translation of the FSMSADF formalism to timed automata (TA). In short, we propose a compositional translation from FSM-SADF to TA that enables computation of some quantitative and qualitative properties of the model not supported by the existing tools, in the UPPAAL model checker. We demonstrate our approach on an MPEG-4 case study which is a typical example of a streaming application from the multi-media domain.