Enhancement in System Schedulability by Controlling Task Releases

In real-time systems fixed priority scheduling techniques are considered superior than the dynamic priority counterparts from implementation perspectives; however the dynamic priority assignments dominate the fixed priority mechanism when it comes to system utilization. Considering this gap, a number of results are added to real-time system literature recently that achieve higher utilization at the cost of tuning task parameters. We further investigate this problem by proposing a novel fixed priority scheduling technique that keeps task parameters intact. The proposed technique favors the lower priority tasks by blocking the release of higher priority tasks without hurting their deadlines. The aforementioned strategy helps in creating some extra space that is utilized by a lower priority task to complete its execution. It is proved that the proposed technique dominates pure preemptive scheduling. Furthermore the results obtained are applied to an example task set which is not schedulable with preemption threshold scheduling and quantum based scheduling but it is schedulable with proposed technique. The analyses show the supremacy of our work over existing fixed priority alternatives from utilization perspective. Keywords—Real-time Systems; Fixed Priority Scheduling; RM Scheduling; Priority Inversion