Scheduling analysis of hybrid real-time task sets

The authors present an integrated scheduling approach which allows to handle real-time task sets with different types of constraints and criticalness. Tasks are scheduled on-line in a dynamic environment based on the earliest deadline first algorithm. Schedulability conditions are derived for hybrid task sets consisting of hard periodic tasks and soft aperiodic activities, interacting through shared resources. Bounded blocking on exclusive resources is achieved by means of a dynamic resource access protocol which also prevents deadlocks and chained blocking. Aperiodic responsiveness is enhanced by an efficient servicing technique which assigns a suitable deadline to each aperiodic request. Moreover the feasibility conditions are extended to handle tasks with deadlines different than periods.