Workload- and Event-Adaptive Dual-Mode Scheduling with Recovery-Aware Service Control for Smart Home Controllers

Abstract

Smart home gateways execute routine automation and emergency workloads on centralized, resource-constrained controllers. Under overload or dynamic workloads, conventional static scheduling policies may allow routine jobs to delay high-criticality tasks. This article presents RT-DASH, a dual-mode scheduling framework for smart home controllers. In Normal Mode, the dispatch is primarily deadline-oriented, whereas Critical Mode, activated by semantic emergency events or sustained workload conditions, gives precedence to ready high-criticality jobs and permits reduction only for application-declared optional computation subject to configured service floors. Once the configured recovery conditions are satisfied, the controller returns to Normal Mode. RT-DASH is evaluated across 15,780 simulation runs against conventional and criticality-aware scheduling baselines. Across the held-out modeled emergency workloads, the recovery-aware RT-DASH variant missed 15.68% of high-criticality deadlines, compared to 62.72% for EDF and 91.67% for period-based fixed-priority scheduling. Furthermore, RT-DASH achieved the same high-criticality deadline miss ratio and mean completed-job response time as the static criticality-first baseline, while completing a larger fraction of routine work. The RT-DASH implementation software and scripts are made public for reproducibility purposes.