SOS: Dynamic Secure Code Offloading for Power Minimization in LEO Satellite Edge Computing

The satellite edge computing (SEC) has recently received considerable attention thanks to its wide area service around the world. However, this also creates a risk of exposing private user data to eavesdroppers. Physical layer security can help prevent this, yet it requires extra usage of network resources. Hence, efficient management of these resources is essential for saving power and ensuring secure code offloading. Moreover, from the perspective of mobile devices that request services, the level of security demands is quite different for various services, yet current studies have not fully considered this aspect. In this paper, we propose a secure code offloading framework for an SEC system with a jamming strategy in the existence of eavesdropping satellite. We formulate an average power minimization problem of an LEO satellite, a gateway, and a mobile device while ensuring security and the stability of queues. This includes making decisions of code offloading, computing/network resource allocation, and jamming unit selection. As a solution of this problem, we propose an SOS algorithm by invoking stochastic optimization theory. Finally, via extensive simulations, we demonstrate that the proposed SOS algorithm can save up to 60% of average power compared to existing algorithms while maintaining the same delay and zero leakage of information toward eavesdropper.