Storage capacity for EDF-ASAP algorithm in energy-harvesting systems with periodic implicit deadline hard real-time tasks

In this paper, a hard real-time system with a single processor that is supplied by a renewable energy source and has a limited capacity storage unit is addressed. Tasks are supposed to be periodic implicit deadline in which execution of tasks must be completed before their deadline and deadline of tasks are at the end of their period. Storage units suffer from leakage moreover power rate of harvesting energy is not precisely predictable. Therefore, scheduling algorithms that are able to schedule tasks as well as consume stored energy as soon as possible are superior in performance. Online EDF-ASAP algorithm that is an energy-constraint-aware derivative of Earliest Deadline First (EDF) is considered in this paper. First, a method to calculate storage unit size for EDF-ASAP algorithm, scheduling energy-harvesting systems is proposed. Another fine storage size adjusting method for low processor utilization task sets is also proposed. Then ability of EDF-ASAP to schedule implicit deadline hard real-time tasks up to processor utilization of 100% and energy utilization of 100% using this fully charged storage unit is proved. Finally, an extensive simulation validates the efficiency of proposed calculation method and compares minimum sufficient storage size with calculated one.