Key points are not available for this paper at this time.
Orbital edge computing (OEC) is gaining interest in multiple application domains due to its potential to mitigate the limitations of traditional bent pipe architecture through in-orbit processing. In one study it was observed that on-board computing facility is idle for about 80% of its mission period. Although processing capability of a single satellite is constrained, but the collective computing power of a constellation is substantial and potentially be exploited for effective distributed in-orbit data processing. In this paper, we propose a distributed scheduling algorithm, DALEOS, for small satellite clusters targeted to meet real-time constraints of data analysis tasks. Our approach efficiently utilizes the constrained resources through workflow partitioning and task distribution. We validated our proposed algorithm for a set of randomly generated task workflow and also on a practical cloud detection use case, in simulation system using satellite tracks generated from datasets downloaded from Celestrak and PlanetLab. The results depict that through efficient utilization of the resources available, the overall latency in task execution can be significantly improved.
Biswas et al. (Mon,) studied this question.