Sharing Resources among Periodic and Aperiodic Tasks with Dynamic Deadlines

In this paper, we address the problem of scheduling hybrid task sets consisting of hard periodic and soft aperiodic tasks that may share resources in exclusive mode in a dynamic environment, where tasks are scheduled based on their deadlines. Bounded blocking on exclusive resources is achieved by means of a dynamic resource access protocol which also prevents deadlocks and chained blocking. A tunable servicing technique is used to improve aperiodic responsiveness in the presence of resource constraints. The schedulability analysis is also extended to the case in which aperiodic deadlines vary at runtime. The results achieved in this paper can also be used for developing adaptive realtime systems, where task deadlines or periods can change to conform to new load conditions.