A Latency-Efficient NF Caching and Offloading Strategy for LEO Satellite Edge Computing

Abstract

With the rapid development of Low Earth Orbit (LEO) satellite networks, computational offloading and Network Function (NF) deployment have emerged as critical enablers for latency-sensitive services. However, existing approaches, such as the GBCO algorithm, typically execute tasks by locating satellites that have cached the required NF across the entire network, resulting in frequent inter-satellite transmissions and consequently increased system latency and energy consumption. To address this issue, this paper proposes a locality-sensitive computational offloading optimization algorithm, termed GCLSO. The proposed method prioritizes task execution on the access satellite to enhance computational locality, and under the condition that cache resources are limited, introduces an NF swapping mechanism to enable adaptive cache adjustment. The algorithm maintains low computational complexity while effectively reducing inter-satellite communication overhead. The simulation results demonstrate that, compared to the GBCO algorithm, the proposed approach achieves a significant improvement in time delay, while maintaining the same time complexity and without any increase in energy consumption.