Don't Blame the Wi-Fi: What I'd Troubleshoot Before Replacing It With 5G

A port in New Zealand replaced its warehouse Wi-Fi with private 5G, and the trade press ran it as a case study. I read it, and I want to be careful here, because there was clearly a real issue with the Wi-Fi for them, and given what they were dealing with it made sense that they went looking for something else. Every site is different, and I obviously don't know everything about this particular installation, so I'm not going to tell you they made the wrong call for their building.

What I can tell you is that I have a feeling the root of the problem could have been found by following a couple of basic troubleshooting steps, and that is the part worth talking about. I'm not saying 5G doesn't work, and I'm not saying it was the wrong decision for them. I'm saying I'm not convinced that 5G fixes a problem the Wi-Fi actually had, and blaming the Wi-Fi for something that's been correctly deployed in tens of thousands of places across the world doesn't seem like the right way to get to the fix.

Basic anomaly detection tells you that if it's working fine for everyone else but not for you, it probably has something to do with the way you're doing it.

So what actually happened at Port Nelson?

The article says they tried a couple of different things, but there wasn't much detail on what anyone did to diagnose what was actually happening from the Wi-Fi side, and there was nothing on how the signal was supposedly being affected by the blue reverse lights on the back of the forklifts. That's not a criticism of the people involved, it's just the piece I'd want to see before I accepted that the Wi-Fi itself was the problem, because a warehouse is a hard RF environment and there are a handful of usual suspects worth ruling out first.

This one is also a little bit unique because it had bottles of wine in it, and those are obviously going to attenuate the signal a lot more than something thinner sitting on a shelf. So there's a real RF problem in that building, I'm not disputing that at all. What I'd want to know is whether it came down to something simple like old equipment, bad antenna cables, or bad antenna placement, because a good site survey goes a long way toward finding exactly that, and none of those things mean the technology failed.

And when a network does struggle, there's a commercial reality around it too. Nobody has really proven out at scale how successful private 5G can be, so when an opportunity comes along, you can understand why a vendor would want to turn a win like this into a story. That doesn't make them wrong, it just means the case study is written from one point of view, and the diagnosis that would tell you whether Wi-Fi ever had a fair shot isn't in it.

Why does warehouse Wi-Fi struggle in the first place?

It's all about propagation in warehouses, and everybody who has deployed one of these knows it. You've got aisles stacked fifteen meters high on both sides, in this case filled with wine, and you have to propagate your signal correctly instead of hoping it fills the room. You've got to use directional antennas that focus the power and the energy directly where it needs to be, because the access points aren't really the important thing here, and what actually matters is the RF energy and the signal and where you direct it.

So instead of using omnidirectional antennas, or 120-degree antennas, or even 60-degree antennas, you want to look at something that gives you 30 to 60 degrees of beamwidth and you want to shoot it right down the corridor toward the clients, the forklifts, and the inventory tagging systems. Directing the energy that way does two things at once, because it saves your client devices from connecting poorly to an access point a couple of aisles away, and it saves your access points from self-interfering and creating co-channel interference with each other.

These are the things we learned really early on doing warehouse and distribution-center deployments, back when people assumed that if they put an antenna up it would work the way it was supposed to, and it turned out they were putting up the wrong antenna. They were reaching for omnidirectional antennas when they should have been reaching for directional ones, and we learned a whole lot about the way access points interfere with each other and reject each other's signal when the antenna choice is wrong. I'm not assuming that's what happened here, but it's the kind of thing a survey surfaces quickly.

What would you check before replacing the network?

A signal propagation assessment you do ahead of time is going to give you a lot of information, and if you have the right software like Hamina or Ekahau, you can go into these buildings and model out what your signal is going to look like. You can do that with such a high level of accuracy now that even if you don't want to perform a full predictive survey, there's a lot you can gather sitting behind a desk with some blueprints and some floor plans before you ever set foot on site.

Now in the event that you really want to see what the attenuation looks like coming off racks filled with wine, then a physical site survey is absolutely warranted, and that's the one place I'd spend the money in a building like this. This isn't a new thing by any means, there are tens of thousands of these networks deployed globally, and the tools have come a really long way, so the honest move is to measure what you actually have and understand where and why it's failing before you decide the whole approach needs to change.

Do this

• Look at your own network first. If you read the Port Nelson story and wonder whether you have the same problem, don't jump straight to 5G. Ask whether you're actually having these issues, and where.
• Run the basic troubleshooting before you conclude anything. Understand what you really have, find where and why it's failing, and zero in on that instead of replacing the whole thing on a hunch.
• Check the cheap stuff early. Old equipment, bad antenna cables, bad connections, and bad antenna placement cause more "Wi-Fi doesn't work" calls than the physics does.
• Look hard at the antennas. If you're running omnidirectional or wide-angle antennas in tall-rack aisles, a switch to 30-to-60-degree directional antennas aimed down the corridor is often the whole fix.
• Model it before you walk it. Use Hamina or Ekahau with your floor plans to predict coverage, then do the physical survey specifically to measure the attenuation your inventory adds.
• Weigh the replacement honestly. If a survey shows the design is sound and it still can't work, that's real information, and a different technology may genuinely be the answer for your site. Just get there by measuring, not by assuming.

Don't blame the Wi-Fi the moment something doesn't work. It's been installed correctly tens of thousands of times around the world, so more often than not it's not the equipment, it's the way that it was installed and engineered, and that's usually a much cheaper thing to fix. Maybe this port is the exception, and if a proper survey said so, I'd respect that. I just wouldn't want anyone reading the headline to skip the survey and assume their own Wi-Fi is the thing to throw out.

FAQ

Is private 5G better than Wi-Fi for a warehouse?
It depends on the building and the diagnosis. Private 5G can absolutely be the right fit in some industrial settings. The caution is only that the Port Nelson story never showed the Wi-Fi being properly diagnosed first, and in most warehouses a correctly designed Wi-Fi network with directional antennas already works. Prove where and why your Wi-Fi is failing before you decide 5G is the fix.

Why does Wi-Fi struggle in tall-rack warehouses?
Propagation. Aisles stacked fifteen meters high, especially with dense inventory like bottled wine, attenuate and reflect the signal. The usual fix is aiming focused RF energy down the corridor with directional antennas rather than blanketing the room with omnidirectional coverage that interferes with itself.

Do I need a full site survey, or can I model it?
Both have a place. Predictive tools like Hamina or Ekahau let you model coverage from blueprints before you visit, which catches most design problems. A physical survey is worth it when you need to measure the real-world attenuation that specific inventory adds.

What usually causes a "failed" warehouse Wi-Fi network?
Design and installation, more often than the gear. Wrong antenna type, bad placement, aging equipment, and bad cabling or connections are the common culprits, and all of them are cheaper to check than a full technology replacement is to buy.

Want this kind of thing weekly, in audio? That's the show.