25GbE SFP28 Fibre Link Planning

1. Confirm the host port

SFP28 describes a module/port family commonly used for 25GbE, but host support, supported optics and firmware remain platform-specific.

2. Do not infer 25GbE from an SFP+ port

SFP+ and SFP28 are related physical families, but a 10GbE-only host does not become 25GbE because an SFP28 module fits mechanically.

3. Choose the optical standard

Select multimode or single-mode optics from distance, fibre plant and supported host modules. Verify wavelength, connector and reach from the exact transceiver datasheet.

4. Check FEC requirements

25GbE links may have platform/optic-specific FEC behaviour. Both ends must agree on the required mode. A mismatch can present as no link or excessive errors.

5. Check patching and polarity

Verify fibre type and connector polish through every panel and patch lead. Inspect and clean before mating.

6. Use DOM and counters

Record optical levels where supported, then watch PCS/FEC/interface error counters under traffic. Link-up alone is not acceptance.

7. Test the path at useful load

Use a controlled throughput test appropriate to the hosts and network. Confirm the rest of the path can actually source/sink the traffic.

8. Document exact part numbers

At 25GbE, generic labels such as 25G optic are insufficient. Record both host ports, optic part numbers, fibre type, patching and FEC settings.

Worked field scenario

A server NIC and switch both have SFP28 cages, but the link stays down with third-party modules. Check the supported optic list, speed configuration and FEC at both ends before assuming the fibre is bad. Mechanical fit is only one part of compatibility.

Procurement questions before ordering

The ordering stage is where many field problems are created. Capture the following before selecting the exact product:

• OS2/OM fibre type and strand count
• Connector and polish at every interface
• Optic wavelength and supported reach
• Expected loss budget and test method
• Patch-panel, splice and polarity plan

Where a manufacturer compatibility list, rail matrix, supported-optics table or battery cartridge reference exists, record the exact source used for the decision. That gives the later installer evidence rather than a verbal assumption.

Fault-finding decision path

Inspect and clean before deeper troubleshooting. Then verify optic recognition, Tx/Rx pairing and polarity. Use DOM as live equipment evidence, OLTS for end-to-end insertion loss and OTDR when event location or characterisation is needed. Keep the design loss budget beside the measurements.

Keep a simple change log while troubleshooting: observation, action, result. If a change does not alter the symptom, put the system back unless there is a documented reason to retain it. This prevents troubleshooting from becoming an uncontrolled configuration drift.

Common mistakes that create repeat faults

Do not look into live fibre. Do not mate APC and UPC interfaces. Do not assume a dust cap means clean, and do not use an OTDR trace alone as a substitute for insertion-loss acceptance when the project requires OLTS testing.

A technically successful installation can still be a poor operational result if the next technician cannot identify what was installed, why it was selected, or how it was tested. Serviceability is part of infrastructure quality.

Acceptance and handover evidence

Provide fibre IDs, route, fibre type, connector/polish, splice locations, optic types and measured results. Store original OLTS/OTDR result files where possible so future technicians can compare traces and losses rather than relying on screenshots.

Use filenames, labels and port/bay identifiers that match the physical installation. Handover evidence should let someone who was not present at the installation verify the final state and repeat the important tests later.