System-Level Evaluation of LEO Satellite Communications Under Service-Driven Traffic Dynamics

As low Earth orbit (LEO) satellite communications take shape on a global scale, system-level evaluation has become essential for a rigorous understanding of their performance characteristics. Most existing studies have relied on a full-buffer assumption, which obscures the heterogeneity and intermittency that characterize actual user equipment (UE) traffic. This article evaluates the performance of LEO satellite communications under realistic service conditions through the incorporation of non-full-buffer traffic models. To this end, we develop a system-level simulator that complies with the channel modeling and evaluation methodologies specified in 3rd Generation Partnership Project (3GPP) technical reports 38.811 and 38.821. Three key system-level performance metrics are considered: UE throughput, resource block (RB) allocation ratio, and packet delay, which are evaluated across 3GPP LEO satellite study cases under diverse traffic models. The throughput results characterize the distribution of achievable data rates, which delineates the practical operating boundaries of the system. The RB allocation analysis reveals patterns of resource consumption as UE density varies, which provides a quantitative basis for assessing system capability. Furthermore, the delay analysis characterizes latency behavior, which is of particular importance in satellite environments where substantial propagation delays are inherent and must be examined to ensure service feasibility. These evaluations provide a realistic performance outlook for non-full-buffer LEO satellite communications and provide insight into the user experience under practical operating conditions.