Infeasibility Test for Fixed-Priority Scheduling on Multiprocessor Platforms

Fixed-priority scheduling (FPS), due to its simplicity to implement, has been one of the most popular scheduling algorithms for real-time embedded systems equipped with multi-processor platforms. While there have been many studies that find sufficient conditions for a given task set to be feasible (schedulable) by FPS with a proper priority assignment, the other direction (i.e., finding infeasible task sets) has not been studied. In this letter, we address a necessary feasibility condition that judges a given task set to be infeasible under FPS with every priority assignment on multiprocessor platforms. To this end, we derive useful properties for the condition and develop the first infeasibility test for FPS on multiprocessor platforms. Via simulations, we show that the proposed infeasibility test discovers a number of FPS-infeasible task sets which are not proven FPS-infeasible by any existing studies.