← All writing Teaching

The case for teaching cloud through real infrastructure, not labs

2 min read

A few of our cybersecurity and cloud students work as student employees in Central Piedmont’s actual data center. They patch real servers, walk down change tickets, sit on real on-call rotations. They are, observably, better at the work than students who only do classroom labs.

The gap isn’t subtle. It’s the kind of gap that makes you stop and ask: is the lab even teaching the thing we think it’s teaching?

What the labs don’t teach

A vSphere lab teaches you to click the right buttons on a vCenter that nobody will ever yell at you about. A real vSphere environment teaches you that there’s a CAB meeting on Thursday, and if your migration window slips past 6 a.m. you’re going to break the morning batch jobs, and the batch jobs have a reputation, and nobody wants their reputation attached to the batch jobs.

Labs teach the verbs. Production teaches the grammar.

That sounds glib but I mean it specifically. In production, you don’t just learn how to do the thing. You learn when to do it, how to tell people you’re going to do it, what happens if you do it wrong, and — most importantly — how to recognize that you should not do it right now. None of that shows up in a lab.

But this doesn’t scale

The honest counterargument is: we can’t put 100 IT-pathway students in the data center. We have eight student employees. Class sizes are 25–30. The model doesn’t generalize.

True. But “doesn’t generalize” doesn’t mean “abandon it.” It means: copy the structural pieces that can generalize.

Three things I’m trying:

1. Lab assignments that mimic real ticket flow. Not “configure this firewall rule.” Instead: “you have a ticket from a clerical user who can’t reach the print server. Here are five facts that may or may not be true. Diagnose and fix. Document for the next tech.” The lab environment is the same. The framing changes everything.

2. Instructors who maintain real systems. Adjuncts and full-time faculty who actually run production for their day job notice things that career instructors don’t. We’ve leaned into hiring industry-current people, even part-time, even at lower-hour commitments than I’d prefer.

3. Rotating crews, not just class sessions. The eight data-center students work on a rotation. Their classmates rotate through too — not as employees, but as “shadow shifts.” Once per term, every student in the program spends an afternoon with the real systems. It’s not enough exposure to change outcomes by itself, but it’s enough to break the lab-only worldview.

The VMware comparison

When I was teaching enterprise techs at VMware, those students were 10× better than my community-college students at the same content. It wasn’t intelligence. It was exposure. Those techs touched production every day. They had context. Every lesson landed against a real system they could already picture.

You can’t replicate that exposure for a first-year community-college student. But you can replicate maybe 1/10 of it. And that 1/10 changes everything.

The cloud and cyber pathways that produce employable graduates next year will be the ones that find ways to get students closer to real consequences. Not pretend consequences. Real ones.

Written by Frazier Smith. Department chair at Central Piedmont, NSF co-PI, and AI & IT educator. More about me →

Subscribe

Get new writing in your inbox.

About one post a month — research notes, classroom field reports, and what's working (or not) in AI + community-college IT education. No spam. Easy unsubscribe.

Prefer RSS? Subscribe via RSS — same posts, no inbox.