An AMF Lab Story: From Bottleneck to Launchpad

‍

An American Tier 1 operator was missing market windows for new 4G and 5G core features because AMF validation in the lab was slow, fragmented, and fully dependent on the vendor for troubleshooting. Multiple lineups, long regression cycles, and limited testing capacity turned every new release into a production risk instead of a growth lever.  

Reailize took over AMF testing as a managed service, treating the lab like a production-critical environment rather than a best-effort shared resource. The team organized the AMF backlog into three clear cycles (regression, new features, and vendor integration) with defined flows from request intake to execution and reporting, so testing stopped competing for time and started running on a predictable rhythm.  

Inside the lab, Reailize handled end-to-end AMF testing operations: preparing configs and environments, creating and refining test cases, executing at scale using Spirent Landslide, and validating results with structured templates so every run generated reusable knowledge instead of one-off reports. When issues appeared, troubleshooting used AMF logs, NAS traces, Qualcomm QXDM, and cross-domain analysis to isolate root causes before involving the vendor, cutting the number of “throw it over the wall” tickets.  

Collaboration changed from reactive escalations to daily, trace-driven joint work with the vendor engineering team and the operator’s own lab and core teams. That included real-time defect follow-up, coordinated change management, support for new 4G/5G interworking releases, cross-domain AMF–MME–ePDG troubleshooting, and enabling testing on new Azure-based infrastructure so the lab kept pace with the operator’s cloud roadmap.  

Over three years, the scope expanded beyond AMF regression into:

• End-to-end and interworking validation across AMF, MME, ePDG, and devices.

• PoCs and dedicated Tier 2 ownership for AMF troubleshooting, with a single, accountable point of contact driving every investigation through to root cause.  

• UE management and testing for major device ecosystems.

• RAN/AMF lab readiness and coordination as new radio and core scenarios came online.  

This model delivered hard outcomes the operator could track release after release: AMF regression cycles were cut by 50 percent, from one month to two weeks, while regression test coverage grew by roughly 65 percent with more base test cases and over 70 feature validations per release. AMF support coverage increased to include 4G MMEs, and RCA cycle time dropped by about 60 percent, from five days to two, turning the lab from a bottleneck into a reliable engine for faster, safer core evolution.

‍