Decentralized satellite federated learning (DSFL) enables LEO satellites to collaboratively train models without relying on terrestrial infrastructures. In DSFL, concurrent model updates for multiple services increase the amount of computation on the satellite, which can lead to longer total training time. This paper proposes a computation-efficient DSFL framework to construct the overlay network for services to minimize the total training time. Simulation results demonstrate that the proposed scheme outperforms the baseline algorithms in terms of total training time.

The rapid changes in network topology caused by the mobility of LEO satellites lead to inter-satellite link (ISL) disconnections. Accordingly, such disconnections will degrade TCP throughput of LEO services. In this paper, we propose a routing algorithm for LEO satellite networks that reduces path changes by avoiding ISLs prone to disconnection. The simulation results demonstrate that the proposed algorithm achieves higher TCP throughput compared to the benchmark methods.