Over the past months, p-NET evaluated Low Latency, Low Loss, Scalable Throughput (L4S) in a 5G Standalone environment, testing performance from no contention to sustained congestion and across different traffic types (TCP and UDP).
While L4S introduces a powerful mechanism for proactive congestion control, our validation revealed a key insight:
No single configuration performs optimally across all network conditions.
Performance depends on factors such as traffic load, traffic type, L4S thresholds, and scheduler prioritization.
For example:
– Threshold tuning improves performance under low congestion;
– Scheduler prioritization becomes critical under sustained congestion;
Key takeaway: To unlock the full potential of L4S, networks must evolve from static configurations to adaptive networks that optimize behavior in real time.
This will be essential to support emerging services such as XR, digital twins, edge AI, and mission-critical systems, and is a step toward intelligent 5G and future 6G networks.
We are now expanding validation across multi-user and real-world scenarios and welcome collaboration with operators, vendors, and partners!