A Novel Embedded Statistical Scheduler for Realising High System-QoS

Development of schedulers for real-time embedded systems is an emerging area of research as it addresses timely scheduling issues like co-operative scheduling, pre-emptive scheduling, micro-kernel based scheduling, etc. Use of statistical schedulers over deterministic schedulers offers the advantage of improved bandwidth utilization, better delay and resource management, higher throughput and thus higher system QoS (quality of service). We report here for the first time a novel stochastic scheduling method based on optimized Markov chains where the pre-scheduled Markov transition probabilities are subsequently dynamically tuned by the central scheduler to give the best system QoS.