A simulation framework for distributed on-board computing for small-satellite clusters

Small satellite clusters are increasingly being used as cost-effective earth observation system. Small satellites, with state-of-the-art sensing payload enable high fidelity imaging, at the cost of high data volume and velocity. Orbital Edge Computing (OEC) was proposed to utilize computing power available on-board to analyze data in-situ to reduce data volume and thereby effectively utilize communication bandwidth. With an extension of this idea od OEC, Distributed Orbital Edge Computing (DOEC) was proposed to collaboratively utilize all computing resource available in a satellite cluster. However, with limited and constrained communication poses challenges for effective utilization of free computation cycles of on-board computers. However, researchers have expressed interest in this technology with the expectation to improve remote sensing capability with better system utilization. Recently, several ideas and proposals around DOEC are presented. Therefore, a standard framework is essential to implement and demonstrate effectiveness of proposals in a realistic setup. In this direction, we present a simulation framework capable of evaluating different performance metrics in a realistic space-like environment. As a demonstration we implement and execute a distributed forest-fire detection workflow on satellite images using the framework.