Capacity-based admission control for mixed periodic and aperiodic real time service processes

摘要

This paper presents an on-line admission control algorithm for one-time aperiodic service requests and periodically generated service processes with real time constraints. Both types of service requests dynamically come and leave the system at run time, making any off-line analysis infeasible. During admission test, periodic task schedulability is determined by using per-task fixed capacity policy to account for the rate of each task. Aperiodic requests are admitted using any remaining capacity after periodic tasks. At run time, earliest deadline first (EDF) scheduling is used to reduce context switch overhead. The proposed admission control is simpler than the traditional EDF analysis for mixed workload of periodic and aperiodic service processes. Simulation results show that the proposed algorithm achieves 10–40% gain in system utilization compared with earlier fixed-priority, bound-based admission test, while incurs only sub-millisecond overhead for each service request.