EPS units on firmware v2.2 sometimes stop answering on the network until their power is cycled. These documents describe the bench investigation (three units, a NETGEAR M4250 switch and a real EPS box), a full code analysis, what was found, the rewritten firmware, and the case for making v3.3 the official firmware — v3.2 plus the gaps the code analysis found, tested overnight against every earlier version.
The decision brief: what the code analysis found, what v3.3 changes, the tested evidence, the cost and compatibility, the honest limits, and the case for adoption. Links into the detail behind each point.
One page · decision brief På dansk · kort orienteringHvad er ændret fra v2.2 til v3.3, og hvorfor vi bør bruge v3.3: anbefaling, de vigtigste tal, forbehold og næste skridt.
Én side · kort På dansk · tabellernev2.2, v3.1, v3.2 og v3.3 side om side: funktioner, fejlmatrix med 17 fejl, genopretningsscenarier, netværksstandarder og daglig drift.
Tabeller For everyone · the tablesv2.2, v3.1, v3.2 and v3.3 side by side: what each does, the full 17-fault matrix, recovery scenarios, network standards, everyday behaviour and resource use.
Tables For decision makers and ITWhat changed from v2.2 to v3.3, how it compares on the bench and against network standards, how it fits enterprise network practice, what we can and can't promise yet, and the rollout plan.
One page For engineersWhat is established, open or excluded; the mechanisms that reproduce the field symptom; charts of the DHCP failure, failure rates with confidence intervals, v2.2 vs v3.0 under load, the standards suite, and the single-register fault matrix (v2.2 / v3.1 / v3.2 / v3.3).
Charts and tables For the recordSetup, methods, every test and result with its evidence grade, corrections to earlier conclusions, limits of the evidence, and next steps.
Long read For the record · previous briefThe earlier decision brief (2026-10-11), kept as it was. v3.3 contains v3.2 unchanged and adds the gap fixes.
One page · superseded by v3.3