A comparison of schedulability analysis methods using state and digraph models for the schedulability analysis of synchronous FSMs

Synchronous reactive models are widely used in the development of embedded software and systems. The schedulability analysis of tasks obtained as the code implementation of synchronous finite state machines (FSMs) can be performed in several ways. One possible option is to leverage the correspondence between the execution of actions in an FSM and the execution of jobs in a digraph task model, thereby applying all the analysis methods developed for these digraph task systems. Another option is to directly leverage the state information and use dynamic programming methods to compute the worst possible sequence of (state dependent) reactions for a given FSM model. In this paper we compare these analysis methods in terms of accuracy and runtime.