Who this checklist is for

I'm a procurement manager at a 150-person network integration firm. I've managed our telecom hardware budget ($340,000 annually) for seven years, negotiated with 25+ OLT and ONT vendors, and logged every order in our cost tracking spreadsheet.

This checklist is for buyer-side procurement staff evaluating EPON or XGS-PON OLT hardware—specifically when you're comparing quotes for 10G OLT units, 2 port EPON OLT boxes, or 2-PON port OLT configurations. It assumes you already know roughly what your network needs. If you're still deciding between PON architectures, sort that out first.

Six steps. Each one has a checkpoint. Skip one and you'll probably find the gap on the invoice, not in the datasheet.

Step 1: Lock your PON architecture before requesting quotes

EPON, GPON, XGS-PON, and 10G EPON are not interchangeable. They run on different standards (IEEE 802.3ah for EPON, ITU-T G.984 for GPON, ITU-T G.9807.1 for XGS-PON), support different split ratios, and behave differently under load. Yet I've received quotes where the vendor mixed terminology inside a single PDF—calling something "10G EPON" in the header and "XGS-PON compatible" in the spec table.

I assumed "2 port EPON OLT" meant two independent PON ports. Didn't verify. Turned out one supplier's "2-port" configuration was one PON port plus one dedicated uplink. That's not a 2-PON device—that's a 1-PON box with a different label.

Checkpoint: Write down the exact standard and port topology you need before you send an RFQ. Require every vendor to quote against that written spec, not their catalog language.

Step 2: Map port count to actual subscriber density

A 2-PON port OLT doesn't tell you how many subscribers it can realistically serve. Port count is only half the equation—split ratio and backplane capacity determine your effective subscriber ceiling.

We had a project where a 2-PON port OLT looked sufficient on paper. Running the numbers with a 1:32 split across both ports gave us 64 downstream connections in theory. In practice, PON port utilization rarely hits 100%, and our traffic model put realistic capacity closer to 48. That was fine for that deployment. It would have been a problem if we'd assumed true 1:64.

More often than not, vendors quote port counts without specifying the split ratio they based their capacity claims on.

Checkpoint: For every port count in a quote, confirm the assumed split ratio. Multiply port count × split ratio × realistic utilization (~70–75%) to get your actual subscriber capacity.

Step 3: Request a TCO breakdown, not just a 10G OLT price

When I started, I compared 10G OLT price quotes line-by-line. The lowest unit price almost always won. That habit cost us real money.

A 10G OLT price sheet typically excludes: power module costs (sometimes sold separately), SFP+ optics (crucially—these can run $40–$200 per port depending on reach and brand), mounting hardware, shipping to your location, and import duties if the vendor ships internationally. One quote we pulled in Q1 2026 for an XGS-PON OLT looked 12% cheaper than the next option. Once we added optics and freight, it was 8% more expensive.

I don't have hard data on industry-wide hidden fee averages, but based on our six years of procurement records, add-on items account for roughly 15–25% of the final invoice on OLT orders where optics aren't bundled.

Checkpoint: Build a simple TCO template before you compare quotes. Include: unit price, power modules, optics per port, mounting, freight, duties, and warranty. If a vendor won't itemize, that's a data point in itself.

Step 4: Verify line card and optics compatibility separately

This is the step most procurement teams skip because it sits between purchasing and engineering. The datasheet says "SFP+ compatible." Great. But which SFP+ modules? At what distances? With what temperature range?

We didn't have a formal verification step for optics compatibility when we started. Cost us when a batch of third-party SFP+ modules worked in the lab at 22°C but dropped links in a cabinet that hit 55°C in August. The OLT was fine. The optics weren't rated for the thermal load.

Checkpoint: Require the vendor to specify tested optic models or at least compatibility criteria (temperature range, distance, wavelength). If they can't, budget for branded optics instead of third-party.

Step 5: Run a pilot order before scaling

Even after choosing a vendor, I kept second-guessing. What if the units behaved differently under load than the samples suggested? The three weeks until the pilot results came back were stressful.

A pilot order is not just about quality. It tests the vendor's lead time, packaging, documentation, and whether the shipping label matches the invoice. I've had a pilot arrive with firmware two versions behind what the quote specified. The hardware was fine; the documentation wasn't.

Checkpoint: Define pass/fail criteria before the pilot ships. Ours are: arrival within quoted lead time ±10%, firmware version matches spec, optics boot on first attempt, and no physical damage. If a pilot fails one criterion, we don't scale up until it's explained.

Step 6: Test the RMA path before you need it

Nobody tests the return process until something breaks. By then, you're negotiating from a position of weakness because your network is already down or degraded.

We didn't have a formal RMA test process. Cost us when a DOA unit took 11 weeks to replace because the vendor's RMA form required a serial number format we didn't have on file. Eleven weeks of workarounds for a $600 module.

Checkpoint: Before your first full order, submit one RMA request for a unit you know works—just to walk through the process. Record response time, required documentation, and whether they ship a replacement before receiving the return. That last point matters more than anything else on this list.

What to watch for

Three mistakes I still see in procurement reviews:

  • Comparing 10G OLT price across different configurations. A 2-port EPON OLT and a 2-PON port OLT are not the same product, even if the labels look similar. Normalize to the same port count, split ratio, and optic package before you compare numbers.
  • Ignoring power budget. It's boring. It's also the reason a "cheap" OLT ends up needing a UPS upgrade you didn't budget for.
  • Mixing refurbished and new quotes in the same spreadsheet. I've done this. It makes the new units look overpriced until you realize the refurb vendor doesn't include a warranty. If you're open to refurbished equipment, keep it in a separate comparison.

I recommend running this checklist against any supplier—LinkScope Sourcing 07 included. A sourcing-focused supplier with 10G and XGS-PON coverage and flexible port configurations is a reasonable starting point for quotes, but that's exactly what it is: a starting point. If you need on-site vendor audits, custom integration testing, or long-term spare parts stocking agreements, a sourcing-focused model may not cover all of it—ask explicitly before you assume.

The honest limitation, in my experience: this checklist works for 80% of standard OLT procurement scenarios. If you're building a greenfield network with non-standard PON requirements or sourcing for a regulated environment where every component needs traceable provenance, you'll need a deeper technical review that a procurement checklist can't replace.

For telecom planning, the article should be read with protocol context in mind: 3GPP TS 38.xxx for radio behavior, IEEE 802.3bt for high-power PoE, ITU-T G.652.D for optical fiber assumptions, insertion loss in dB for link budget, and PIM in dBc for passive RF quality.