Energy-Aware Scheduling and Dynamic Reconﬁguration in Real-Time Systems
Various research eﬀorts have focused on design and optimization of real-time systems in recent years. These systems requireuniquedesignconsiderations since timingconstraints are imposedon theworkloads (i.e., tasks). Tasks have to complete execution by their deadlines in order to ensure correct system behavior. For hard real-time systems, as in safety-critical applications like medical devices and aircrafts, violating task deadlines may lead to catastrophic consequences. Due to these stringent constraints, realtime scheduler must perform task schedulability analysis based on task characteristics such as priorities, periods, and deadlines . A task set is considered to be schedulable only if there exists a valid schedule that satisﬁes all task deadlines. As embedded systems become ubiquitous, real-time systems with soft timing constraints, in which missing certain deadlines are acceptable, gain widespread applications
including gaming, housekeeping, as well as multimedia equipments. Minor deadline misses may result in temporary service or quality degradation, but will not lead to incorrect system behavior. For example, users of video-streaming on mobile devices can tolerate occasional jitters caused by frame droppings, which does not aﬀect the quality of service.