OEC-H2O: A Hierarchical Framework for Joint Service Orchestration and Computation Offloading in Orbital Edge Computing Networks

Orbital edge computing (OEC) is expected to provide orbital data processing, conserving communication resources and enabling real-time computing responses. In view of the uneven computing resources and computation task requests of low earth orbit (LEO) satellites, computation offloading is widely considered to optimize resource utilization of OEC networks and improve computing performance of tasks. During this process, how diverse computing services are orchestrated plays a non-negligible role in the performance of computation offloading. Therefore, this article introduces a hierarchical framework for joint service orchestration and computation offloading in OEC networks (OEC-H2O). In this framework, computation offloading is addressed on small timescales, using a Markov decision process (MDP) to optimize task processing delay, energy consumption, load balancing, and packet loss. Considering that frequent orchestration of services will bring huge orchestration overhead, service orchestration is optimized on larger timescales, which is modeled as an integer programming problem to enhance computing service capabilities of OEC networks and reduce orchestration costs. A hierarchical approach employing deep reinforcement learning (DRL) and heuristic algorithms aims to iteratively reveal optimal solutions. Extensive simulations are conducted to verify the effectiveness and superiority of the proposed scheme. Finally, we discuss potential open issues and directions for future research.