Inter-Satellite Links-Enabled Cooperative Edge Computing in Satellite-Terrestrial Networks

Satellite edge computing (SEC) ignites ubiquitous computation offloading across the Earth’s surface. However, non-uniformly distributed users generate uneven traffic demands, resulting in excessive load on particular satellites and insufficient use of satellite computing resources. In this paper, we explore the possibility of cooperative SEC using inter-satellite links, where we design a two-stage cooperative SEC algorithm in satellite-terrestrial networks. Specifically, in stage one, users can offload their partial computational tasks to their directly connected satellites and locally execute the rest of the tasks. In stage two, the connected satellites further offload partial computational tasks to the terrestrial station and available satellites in a cooperative manner, especially for high-traffic satellites. We aim to minimize the system weighted-sum energy consumption during SEC by jointly optimizing task allocation, power control, bandwidth allocation, task partition, and computing resource allocation under the constraints of maximum tolerated latency, computation capacity at each satellite, total bandwidth, and maximum allowable transmission power. Furthermore, we introduce an iterative algorithm by decomposing the original non-convex problem into several sub-problems and solve each sub-problem with attempts to derive theoretical analysis. Simulation results demonstrate that our introduced cooperative SEC algorithm can efficiently use satellite computing resources and significantly reduce system weighted-sum energy consumption compared to other algorithms.