SEQD: An Optimized Approach for Microservice Function Chain Placement in Satellite Edge Clouds
Yunlong Liu, Xiaojie Ju, Renchao Xie, Qinqin Tang, Tianjiao Chen, Tao Huang · 2024
In the era of sixth-generation (6G) networks, satellite edge clouds have become essential for delivering globally accessible, low-latency services. However, deploying microservices within these networks presents significant challenges due to resource limitations, high latency, and the dynamic nature of satellite environments. This paper addresses the optimization problem of microservice function chain placement in satellite edge clouds, with a dual objective of minimizing end-to-end latency and placement costs. To tackle this, we propose the Q-Learning Enhanced Drops in Satellite Edge Clouds(SEQD) algorithm, which integrates reinforcement learning with heuristic optimization to achieve efficient and adaptive microservice placement. Simulation results demonstrate that SEQD achieves faster convergence and greater latency reduction compared to baseline algorithms, particularly in large-scale network scenarios, showcasing its potential for enhancing resource utilization and performance in satellite edge networks.