I've spent the better part of the last four years reviewing deliverables for a major networking equipment provider. My job is to ensure that every switch, access point, and configuration guide that leaves our hands meets a specific standard before it reaches a customer. It’s not a glamorous role. It’s about catching the small stuff—a typo in a CLI command, a misleading spec sheet, a tolerancing error in a mounting bracket—that can turn a routine deployment into a $22,000 redo.
And that experience has fundamentally changed my view on enterprise networking. The old adage, 'Just make it work,' is no longer sufficient. It's not just outdated; it's a liability. We're past the point where plugging a switch into a rack and hoping for the best is an acceptable strategy. Here's why I believe the biggest risk to your network infrastructure is treating it like a commodity.
The Myth of 'Set It and Forget It'
There's a persistent myth in IT that once a network is up and stable, the work is done. You set the VLANs, configure a few routing protocols, and walk away. This was vaguely true ten years ago when the primary goal was just moving data from point A to point B. Traffic patterns were predictable. Security was a perimeter problem. You had a main office and maybe a few remote sites.
Today, that's a fantasy. Your network is no longer a simple pipe; it's a living organism. You've got IoT sensors, guest Wi-Fi, video conferencing, cloud applications, and a workforce that’s distributed across dozens of locations. The old approach to segmentation—flat VLANs and ACLs—simply can't handle the complexity. It's tempting to think you can just add more bandwidth and it will fix everything. But a 10Gb link won't help if your network can't dynamically isolate a compromised IoT device or prioritize Zoom traffic over a firmware download. The 'more bandwidth' advice ignores the fundamental shift from a transport network to an application-aware platform.
The 'Fast vs. Quality' Tradeoff is a Trap
I see this all the time. A project manager is under pressure to get a new site online. They have a limited budget and a tight timeline. The default move is to spec hardware based on port count and price, configure it with a template from the last project, and move on. It’s fast. It’s cheap. And it usually works... for about six months.
I had a specific incident last year that cemented this. A partner was rolling out a new SD-WAN deployment for a retail client. They opted for a lower-spec router because the throughput numbers looked fine on paper. They skipped the advanced path selection configuration because 'the main link is stable.' It was working fine during the pilot. Then Black Friday hit. The moment traffic spiked, the router choked. The main link didn't fail, but the burst traffic overwhelmed the hardware. They had 2 hours to decide whether to pull the pilot unit and delay the rollout (costing $15k in penalties) or swap the hardware on-site (costing $4k in labor).
In hindsight, they should have invested the extra $600 upfront for the right hardware and spent the two hours tuning the SD-WAN policies. But with the CFO watching the budget, they made the call with incomplete information. That’s the trap. The 'fast' decision feels responsible in the moment, but it forces you into a reactionary mode later. Quality isn't a cost; it's an insurance policy against the inevitable 'Black Friday' moment.
Where ‘Good Enough’ Fails: Wi-Fi 6E and Security
Let’s talk about edge technologies. I’ve reviewed dozens of proposals where the spec says '802.11ax APs' without specifying the band configuration. The assumption is that all Wi-Fi 6 is the same. It’s not. If you're deploying for high-density environments—think classrooms, conference halls, or manufacturing floors—sticking with only the 5GHz band is a self-inflicted bottleneck.
Wi-Fi 6E, with its dedicated 6GHz spectrum, isn't just a speed bump. It’s about channel width and interference reduction. On a standard 5GHz deployment, you might have three or four non-overlapping 80MHz channels. On 6GHz, you have fourteen. That’s a fundamental difference in capacity planning. A company that skips on Wi-Fi 6E today isn't saving money; they're limiting their ability to run high-bandwidth, low-latency applications tomorrow.
And then there's the security side. The 'good enough' thinking on network segmentation is dangerous. A lot of organizations still rely on flat network topologies because 'we have a firewall at the edge.' That was true in an era before ransomware could move laterally in minutes. Fabric Connect technology (like what Extreme Networks offers) allows for micro-segmentation at Layer 2 without complex ACLs or overlays. Earlier this year, a manufacturing client had a zero-day vulnerability hit their HVAC controller. In a traditional flat network, that controller would have had unfettered access to the production network. But because they had implemented a fabric-based network with automated endpoint classification, the compromised controller was automatically isolated. The attack was contained before it reached any critical systems. The 'simple' VLAN solution wouldn't have caught that.
The Bottom Line: Plan for the Problems You Haven't Seen Yet
I know the pushback. “We’ve been running this way for years and it’s fine.” Or, “We don’t have the budget to over-engineer the network.” I get it. Budgets are real. But I’d argue that the most expensive network is the one that requires a complete redesign in 18 months because you didn’t account for IoT or SD-WAN or high-density Wi-Fi.
I went from someone who focused purely on hardware specs to someone who looks first at the operational intent. What is the network supposed to do in two years? Can it isolate a guest device from a building management system? Can it automatically route critical traffic over the best link? Can it handle a 3x increase in wireless clients? If the spec sheet doesn’t answer those questions, it’s not a good network—it’s just a bunch of expensive boxes.
The fundamentals of a good network haven't changed: it must be reliable, secure, and manageable. But the execution has transformed. The 'just make it work' era is over because the stakes are higher. The cost of a bad network isn't just a slow application; it's a production line that stops, a sales call that drops, or a security incident that costs millions. We have to stop buying networks like they are commodities. We need to start designing them as strategic assets.