On-line schedulers for pinwheel tasks using the time-driven approach

Pinwheel scheduling algorithms can be used to produce distance-constrained real-time system schedules where the temporal distance between any two consecutive completions of a task must be less than a pre-defined time interval. A pinwheel schedule can be generated off-line and executed cyclically. Such an approach provides a good predictability and allows for off-line schedule optimization. However, the static approach is inflexible and may require a large space to store the schedule. By taking advantage of the harmonic property between pinwheel task periods, one can generate the pinwheel schedule dynamically at run time in polynomial time and space. In this way, efficient and flexible time-driven schedulers can be implemented. The authors show the algorithms and study the practical issues on implementing on-line pinwheel schedulers.