Effective Capacity Analysis for Integrated Communication and Computation for LEO Satellite Edge Computing

Low earth orbit (LEO) satellite networks integrated with edge computing capabilities offer an efficient and flexible task offloading solution for ground users over wide areas, making them a promising approach to realizing the ubiquitous computing vision in 6G. However, in dynamic LEO satellite networks, the analysis of system capacity based on integrated communication and computing queues has not been sufficiently studied. This paper investigates the performance of a two-hop satellite edge computing system from the perspective of communication-computation integration. Considering the mobility of satellites, the transmitter is assumed to have incomplete instantaneous channel state information of the satellite. Based on this, we derive the effective capacity expressions for a two-hop (ground-to-satellite link and inter-satellite link) satellite system with joint communication and computation queues over Nakagami-m channels. The probability expressions that the cascaded communication and computation queues do not violate the delay bound are derived. Moreover, the achievable arrival rate bound of the two-hop satellite edge computing system with limited communication and computing resources is analyzed in detail. The simulation results show that 1) The trade-off between communication and computing resources dominates the capacity of the two-hop satellite edge computing system; 2) Proper resource allocation is necessary to balance communication and computation, avoiding bottlenecks and maximizing system throughput.