Main Results on Earliest Deadline First Scheduling for Energy Neutral Sensors

Abstract

Energy harvesting technology permits perpetual operation to small electronic devices such as wireless sensors. Intermittent computing and real-time scheduling are the two main issues when designing such devices. Classical scheduling approaches are not convenient because they cannot accommodate fluctuations in power supply. A variant of the Earliest Deadline First scheduler was proposed so as to adapt to the characteristics of energy harvesting computing systems. The so-called ED-H algorithm is an idling and clairvoyant one which improves performance in terms of deadline missing ratio. And the so-called SSP algorithm is an aperiodic task server which provides optimal responsiveness for aperiodic tasks while guaranteeing deadlines of the critical periodic tasks. Thus, this paper reviews the main advances in real-time scheduling theory for energy harvesing systems under real-time constraints.