Scheduling distributed safety critical systems

A systems architecture is presented that will guarantee crucial timing requirements, is tolerant of a range of hardware errors, but is sufficiently flexible for sporadic activities to be guaranteed and nonreal-time tasks to be accommodated. The model presented uses static allocation of application tasks, preemptive priority based scheduling, and replication for task availability. The architecture proposed has been obtained by choosing (known) fault tolerant techniques that are amenable to timing analysis, extending schedulability models to incorporate these techniques, and using, to advantage, properties of the computational model usually applied in hard real-time applications