Ericsson and CSPs implement L4S standard to cut network latency
Ericsson details the Low Latency, Low Loss and Scalable throughput (L4S) framework, a standard adopted by IETF, 3GPP, and DOCSIS, designed to provide consistent, low-latency experiences in congested networks. L4S uses early congestion signaling to help applications adapt before performance degrades, particularly benefiting interactive video, cloud gaming, and XR services. Several CSPs, including SoftBank and MasOrange, have implemented L4S to improve real-time service responsiveness on their 5G Standalone networks.
Key Takeaways
- SoftBank achieved a 90% latency reduction by combining L4S with 5G Advanced on its live Standalone network.
- MasOrange has deployed L4S across 20 major Spanish cities, including Madrid and Barcelona, to improve real-time video and AR/VR responsiveness.
- The L4S framework is a cross-industry standard adopted by the IETF, 3GPP, and DOCSIS for both mobile and cable access.
- L4S utilizes Explicit Congestion Notification (ECN) marking to signal endpoints to adjust sending rates before network buffers overflow.
- Ericsson’s solution integrates L4S across its 5G RAN, Core, and Transport portfolios to enable "differentiated connectivity" based on experience rather than peak speed.
Why It Matters
L4S represents a critical technical shift from speed-based to experience-based network service levels. By enabling a real-time feedback loop between the network and the application, L4S allows streaming services to degrade quality gracefully via resolution or frame rate adjustments rather than suffering the abrupt freezes typical of best-effort delivery. For the broader ecosystem, this standardizes how high-bandwidth, latency-sensitive applications like 4K cloud gaming and XR operate across fragmented network environments. As more CSPs move toward 5G Standalone, L4S will likely become a prerequisite for premium, jitter-free service tiers. Watch for individual streaming app integrations, as consumer-side performance gains depend on both the network and the client-side player supporting the ECN signaling.
Additional Context
The commercial momentum for L4S has accelerated as major hardware and software providers integrate the standard into consumer ecosystems. Per Apple, June 2023, the company added L4S support to the networking stacks of iOS 17, macOS Sonoma, and tvOS 17, specifically targeting improvements for FaceTime quality. This device-side readiness is essential because L4S requires both the network bottleneck and the endpoint application to recognize and act upon congestion markings. In the broadband sector, Comcast began field trials of Low Latency DOCSIS (LLD) in 2023, collaborating with Apple, NVIDIA, and Valve to optimize performance for the Xfinity 10G network. According to RCR Wireless, January 2025, Comcast's commercial rollout has expanded to cities including Atlanta and San Francisco, positioning lower lag as a primary competitive differentiator over 5G home internet. Gaming platforms have also emerged as early adopters to solve the high-bandwidth, low-latency paradox. Per NVIDIA, December 2024, the GeForce NOW cloud gaming service officially introduced L4S support, allowing Ultimate tier subscribers to monitor the feature’s status via in-stream statistics. This adoption aligns with findings from the Broadband Forum in February 2026, which demonstrated that L4S-enabled traffic maintains stable round-trip delays of roughly 20 milliseconds even under heavy congestion, compared to 35 milliseconds or higher for classic traffic. As network operators like MasOrange undergo structural changes—including its 100% acquisition by Orange group in June 2026—the deployment of 5G Advanced and L4S is increasingly viewed as a prerequisite for supporting the massive traffic growth predicted for AI and real-time interactive services.
Read full article at ericsson.com
Get this in your inbox → Subscribe
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source