This research compares the performance of two widely used open-source 5G core (5GC) platforms, Free5GC and Open5GS, in a resource-constrained private network environment. While previous studies have mainly focused on feature comparison or large-scale deployments, performance under limited computational resources has received less attention, particularly for small-scale enterprise use cases. In this work, both platforms are integrated with UERANSIM to emulate end-to-end 5G communication and evaluated under dynamic user equipment (UE) scaling. Each 5GC instance and simulator component is allocated one CPU core and 2 GB of memory. Performance is assessed using key Quality of Service (QoS) metrics, including throughput, latency, packet loss, and resource utilization (CPU and memory), under both TCP and UDP traffic. The results show that Open5GS consistently provides better performance than Free5GC. It achieves up to 10.58 Mbps throughput compared to 9.22 Mbps and maintains lower latency around 0.72–0.73 ms, while Free5GC reaches up to 1.20 ms as the number of UEs increases. In addition, Free5GC reaches high CPU utilization earlier under increasing load. These differences are mainly related to its microservice-based architecture, which introduces additional processing overhead.
Copyrights © 2026