Finisar Transceiver Guide: A 5-Step Buying Checklist from an Admin Who Learned the Hard Way

I'm an office administrator at a mid-sized company, which means I'm the one who gets asked to "just order some transceivers" by the network team. I've been handling our optics purchases since 2020—about 40 to 60 line items a year, from SFP+ to QSFP28, with a budget that my finance director watches closely. I'm not an engineer. But I've made enough expensive mistakes to know what actually matters when you're buying Finisar optical modules.

This checklist is for buyers like me: people who need to order the right 100G transceiver without a deep degree in photonics. It's not a substitute for your network team's judgment. It's a way to keep you from buying the wrong thing (or overpaying for the right thing) in the first place.

Here's the five-step version. Do these in order.

Step 1: Verify Compatibility Down to the Firmware Level

The first thing everyone checks is the form factor. A 100G QSFP28 module is a 100G QSFP28 module, right? Not exactly. The port might look the same, but the switch's firmware can reject a perfectly good module if it isn't on the vendor's compatible list.

I discovered this in our 2023 network refresh. We had a quote for Finisar 100G modules from a distributor, and a cheaper "compatible" option from another reseller. The cheaper ones did not link up on our Cisco Nexus 93180YC. Turns out the switch's IOS version at the time didn't include that third-party vendor's part number in its transceiver matrix. The Finisar modules were on the list.

So the checklist here is:

• Get the exact switch model and the running firmware version.
• Look up the vendor's "Qualified Optical Transceivers" document. Cisco and HPE both publish these.
• Confirm that the specific Finisar part number is listed for that firmware.

Don't skip this because a salesperson tells you "it'll work." They might be right. But "probably" costs money.

Step 2: Match the Module to Your Fiber and Distance

This is where the "value" logic gets interesting. A 100G-SR4 module for multimode fiber might cost $300. A 100G-LR4 module for singlemode fiber might cost $700. They look almost identical from the outside, but they're not interchangeable.

We had a candidate vendor push us to buy LR4 modules because "they're better." Not true—they're different. Our short, rack-to-rack links were multimode OM4. SR4 was the right choice for that segment. The LR4s would have been a waste of money, and not just because of the higher upfront price. They also need singlemode fiber with LC connectors, while the SR4 uses MPO/MTP. We'd have had to re-cable the racks.

So before you compare prices, check two things:

• What type of fiber is in the path? Look at the cable jacket: OM3/OM4 means multimode; OS1/OS2 means singlemode.
• How far is the actual link? 100m, 2km, 10km? That determines SR4, LR4, or ER4.

I know this sounds basic. But I've seen purchase orders for a whole batch of modules that were right for the switch port and wrong for the fiber.

Step 3: Calculate the Optical Link Budget—It's Not That Hard

Here's the step most people skip, and it's the one that saves you from "why does this link keep dropping at 1 a.m.?"

The 100G Ethernet PHY types (SR4, LR4, ER4) are defined in IEEE 802.3ba. Your switch's datasheet will tell you which PHY types it supports. The transceiver must match.

Every transceiver datasheet also lists a range for transmitter output power and the receiver sensitivity. Subtract the minimum transmitter power from the receiver sensitivity, and you get the worst-case link budget.

As an example, take a Finisar 100G-LR4 QSFP28 (the FTLX1371D3BCL). The datasheet says the per-lane transmitter power range is -2.5 to +3.5 dBm, and the receiver sensitivity is -10 dBm. The absolute worst-case budget is -10 minus -2.5 = 7.5 dB. That's your worst-case number. If your fiber path has connectors, patch panels, and splices, you can measure the loss with a simple optical power meter. You want at least 2 dB of margin above the calculated loss.

Our network team taught me this after a sub-procurement earlier. We'd bought modules that "technically" fit our distance, but the link had excessive loss because of poor fusion splices. The module's built-in DOM showed high bit error rates. That's not the module's fault—it's physics.

I'm not saying you should become a fiber-optic engineer. But you can ask the network team for the link loss numbers. If they don't have them, that's a question worth asking.

Step 4: Price Isn't the Whole Story—Weigh the Total Cost

Let's talk money, since that's usually the real driver of a buying decision. The cheapest 100G module you can find might be $350. A genuine Finisar might be $700. You see the price difference and think "I can save enough to pay for a decent lunch." But that's the upfront price, not the total cost.

What's the total cost? Here are the line items that never show up in a purchase order:

• If the cheap module fails, will the switch vendor support you when you call about a port issue? Many switch vendors decline to troubleshoot if they detect third-party optics in the logs.
• If the module throws a DOM sensor reading that's garbage, how many hours will the network engineer spend debugging before swapping it out?
• If the module isn't authentic or is a "re-labeled" unit, what's the cost of the downtime when it dies? Our facility generates about $4,000 an hour in revenue, so a 30-minute outage pays for the difference between cheap and genuine.

I had this exact lesson in 2021. I ordered a batch of ultra-cheap SFP+ modules from an online marketplace. They worked for two weeks. Then one started flapping, and we discovered the "vendor code" in the switch was unrecognized. We swapped in a Finisar module we had in stock, and the link was stable. The "savings" disappeared in the cost of a rush shipment and a wasted morning.

So my advice: buy from an authorized Finisar distributor. Ask about warranty and return policy. Make sure the modules come in sealed anti-static packaging with the correct label. Counterfeit transceivers are common enough that checking the manufacturer string in the switch CLI (it should read "Finisar Inc." or the current Coherent brand) is worth doing.

Step 5: Test on the Real Hardware Before Deployment

You've got the modules in hand. Don't rack up all of them and expect the works. Test one module in the exact switch model, with the exact firmware, and the exact fiber patch cable you'll actually use. Let it run for 24 hours and check the optical stats.

I learned this when we deployed 16 Finisar 100G modules across two switches. The first switch accepted the module instantly. The second switch, with a different line card and a slightly older firmware version, would not link. If I'd tested one module on both switches before ordering in bulk, we'd have updated the firmware first instead of scrambling.

So the test step is simple:

• Install one module in a live port (or a lab port if you can).
• Verify the link comes up and stays up.
• Run "show interface transceiver" (or equivalent) to check the module's vendor ID, temperature, and optical power.
• Repeat for each distinct switch model and firmware combination you're targeting.

This isn't a guarantee you'll avoid every issue. But it'll catch 80% of the compatibility problems before you've opened 24 boxes.

Common Mistakes and Things to Keep in Mind

Here are a few extra notes from real purchase orders:

• Don't assume "100G" means the same on every port. Some switches use breakout cables to carry 4×25G. That's a different module than a single 100G link in many cases.
• Check whether the switch supports the module's digital diagnostics monitoring (DOM). If not, you lose visibility into optical power. Finisar modules support DOM, but the switch must too.
• Finisar is now part of Coherent (it was acquired by II-VI in 2019, and II-VI became Coherent). So don't be confused by the branding change. The know-how and product lines continue.

One more caution: don't buy more than you need. We've all been in the position where the distributor offers a volume discount on 50 units when you only need 12. That 38-module buffer might be useful if the network grows. But transceiver prices drop over time. In November 2024, we bought Finisar 100G-LR4 QSFP28 modules for about $680 each at volume. If we'd bought a large stock in 2022, we would have overpaid by a third.

So, do the homework. Verify compatibility, match the optics to the fiber, calculate the budget, weigh the total cost over price, and test before deploying. That's the checklist. It's not fancy, but it's kept my CFO off my back and my network stable.