Service Offloading in Terrestrial-Satellite Systems: User Preference and Network Utility

In this paper, we investigate service offloading in an integrated terrestrial-satellite (T-S) system. We consider the terrestrial base station (TBS) and satellites to be service providers, all user equipment (UE) to be service requesters, and the service can be content delivery, computation, etc. While offloading services to the satellites can prevent the TBS from being overloaded, the quality of service (QoS), e.g., content delivery latency, may degrade, necessitating a balance between the user preference and the utilities of the TBS and satellites. From the perspective of network management, we propose an abstract model that incorporates the utilities of the TBS, the satellites, and the UE, as well as the service capacity, service load, and service cost at the TBS and the satellites. While finding the optimal offloading decision, a problem of integer programming, is NP-hard, we develop two algorithms with low complexity for finding sub-optimal solutions of the offloading decision problem in the scenarios of one satellite and multiple satellites, respectively. Moreover, we prove that the solution found by the first algorithm is guaranteed to be optimal under the condition that the tasks for service from all UE have an identical size. Numerical results demonstrate the performance of the proposed algorithms compared to that of the optimal offloading by exhaustive search and the offloading by the greedy algorithm.