I Don't Trust Networks That Weren't Tested Before Commissioning
Here's something that might sound contrary to the "move fast" philosophy: I'd rather hold up a deployment by a morning to test the fiber than risk a week of troubleshooting later. I'm the quality inspector who reads the OTDR trace before the switch team even unboxes the gear. And yes, I've been called overly cautious.
But after reviewing over 200 fiber plant acceptance reports annually for the past four years, I've stopped apologizing for it. Most network issues—maybe 80% in my experience—are physical-layer problems that were present before the switch or router was connected. The light meter doesn't lie, and the visual fault locator will show you exactly where the problem is. The question is whether you look before or after the outage.
Argument 1: The Hidden Fault That Looks Good on Paper
From the outside, it looks like a new fiber run is either "good" or "bad." The reality is more nuanced. I've seen plenty of links that pass a basic continuity check but fail under real traffic. A quick test with an EXFO light meter—say, the FIP-400B series—will show you actual insertion loss. But a lot of teams skip this step because the cable was "certified" by the installer.
In Q1 2024, we received a batch of 12 pre-terminated trunk cables for a data center expansion. The vendor's documentation showed they all passed. But when we spot-checked with our own meter, three of the twelve were 0.8 dB over spec on the far end. Normal tolerance for that run was ±0.3 dB. The vendor claimed it was "within industry standard." We rejected the batch, and they redid it at their cost. Now every contract includes a clause that we reserve the right to verify with our own equipment.
That 0.8 dB wouldn't cause a link failure immediately. But in a DWDM system with 40 channels, that extra loss accumulates. It might cause intermittent errors months down the line, when you've already forgotten which cable was which. The EXFO visual fault locator—the 2660 flip, specifically—made the faulty connector visible in about 30 seconds. Hard to argue with a red glow at the splice.
Argument 2: The Real Cost Isn't the Test—It's the Repeat Truck Roll
Most buyers focus on the upfront cost of testing equipment and completely miss the cost of not having it. A field technician costs around $75–150 per hour depending on the market. If they have to roll out twice—once to install, once to troubleshoot—you've just doubled your labor cost for that site. That's before you factor in the switch that's sitting idle in the rack.
I'll be honest: I wasn't always this obsessive about pre-testing. Early in my career, I figured the switch's built-in diagnostics would catch any problems. Then we had a fiber cut—not a dramatic one, just a micro-bend at a patch panel—that caused a 0.5 dB loss on a 10G link. The switch reported it as "link up." Traffic passed. But the bit error rate was high enough to cause application timeouts. It took us three visits to find it. First visit: "Link looks fine." Second visit: "Maybe it's the switch." Third visit: "Oh, it's the fiber."
The most frustrating part: a 30-second test with a PON power meter on the first visit would have revealed the problem immediately. That experience cost us about $2,200 in labor and delayed our go-live by a week. Since then, I've made it standard practice to test before connecting any active equipment.
Argument 3: The "Cisco Switch vs. EXFO Gear" False Choice
I've heard this argument more times than I can count: "We spent a lot on the Cisco switches. The fiber should be fine." I get the logic—people assume good gear upstream means everything downstream is okay. That's exactly wrong.
Think of it this way: the switch and the cable are different layers. A top-tier Cisco switch (or any switch, really) will operate perfectly on a bad cable—operating, but erroring. The switch can't fix a dirty connector or a bad splice. It can only report what it sees. The EXFO light meter and visual fault locator are looking at the medium itself, not at the endpoint. They're testing the stage before the actors arrive.
I ran a blind test with our deployment team: same network segment, with and without a pre-test using the 2660 flip and a basic power meter. The group that tested first had zero post-commissioning issues in that segment. The group that didn't? They had two tickets within the first month—both physical-layer problems. The cost difference wasn't the equipment; it was the time spent guessing.
Now, I'll concede that some people are skeptical. A colleague once told me, "But the vendor certified the cable. Why are we double-checking?" To be fair, most vendor certifications are legitimate. The problem is that certification a month ago doesn't guarantee the cable wasn't damaged during installation. A fiber that tested fine on the reel can get kinked when pulled through a conduit. A connector that was clean at the factory can get dirty when terminated in the field. The question isn't whether the vendor is trustworthy; it's whether the final installed state matches the spec.
Final Recommendation: Test Before You Connect
Honestly, I'm not sure why this practice isn't universal. My best guess is it comes down to schedule pressure—people feel they can't spare 15 minutes to test when the project plan is already tight. But I've seen that 15-minute delay save two days of rework often enough to believe it's almost always worth it. The EXFO FTB-1v2 or a simple handheld meter like the FOT-100 is all you need. You don't need a full lab setup. Just a quick insertion loss and a visual inspection.
I get why people want to connect the switch first and "see if it works." It feels faster. But network reliability is rarely about the first day; it's about the first month, the first quarter, the first year. A cable that just barely passes on day one will get worse over time as temperature cycles and physical movement take their toll. The hidden problem is that you've introduced a latent defect into your network. The only question is when it will become visible.
This was accurate as of early 2025. The test equipment market changes fast—EXFO has updated their FIP series a couple of times—so verify current models and pricing before budgeting. But the principle hasn't changed: measure the medium before you activate the gear. It's the cheapest insurance you'll buy.
