As a network engineer, the best way to get better at the craft is to practice it, and I thought I was on the right road to success : I do my reading, finish my labs, configure the interfaces just like that guy from youtube asks me to. With all this “experience”, I find myself often stuck when a client asks me to do something on his network, after a year of working, not one single person asked me to configure vlan 10 on the GI0/0 interface. Instead I am faced with non-technical people demanding solutions to problems they can hardly describe. And after thinking about it, I don’t really blame them, after all the network was made to be working, and if you’re not really interested in the field, there is no reason for you to know how it works, the most important thing for you is that it does (which is kind of sad, because no one thanks you when there is no problem). This led me to the conclusion that we don’t only require the knowledge to acheive the needs of network, but also the translation skill to figure out what are these needs. And I felt like a lab with direct questions, and answers inside those questions is not the way to cut it.

So I did some reverse engineering.

An RFP for a hospital that doesn’t exist⌗

If you can’t find a client that would hire you, create one.

In consulting, when an organization wants a network built, they don’t send you a task list. They send you a Request for Proposal. A document that describes their business, their sites, their problems, and what they need. Then it’s your job to design the answer.

I wrote one of those for myself.

Valkyrion is a regional healthcare group. It has a headquarters and three hospitals. Its current network is flat and improvised: clinical systems share the wire with guest wifi, medical devices sit next to staff laptops, nothing can be isolated when something goes wrong, and one failed internet box at a site cuts doctors off from patient records.

None of it is real. All of it is the kind of thing that is real.

Then I hired myself to fix it.

Why hospitals ?⌗

Other than the dramatic effect and feeling like being casted in grey’s anatomy, I picked healthcate because it makes every design decision have an obvious reason.

In a normal lab, “put the guest network on its own VLAN” is a step. In a hospital, it’s a sentence you could say to a room full of people who are not engineers: patients and visitors should not be on the same network as the hospital staff. Segmentation stops being a switch feature and becomes patient safety.

Same with redundancy. “Add a second core switch” is a step. “A doctor should not lose access to a patient’s chart because one box died” is a reason.

Three sites, three shapes⌗

The three hospitals are deliberately different sizes.

Alpha is the biggest. Full three-tier design: access, distribution, core. Beta is mid-size. Collapsed core, two tiers. Gamma is the smallest and newest. A SOHO-style site, one combined footprint.

The CCNA asks you to compare two-tier, three-tier and SOHO topologies. Most people learn that from a diagram in a slide. I wanted to build all three, in one network, and feel why you would choose each.

The headquarters holds the shared services everything else depends on : records, DNS, DHCP, NTP, logging. Which means when you break HQ, you break all four sites.

The hardest part⌗

I expected the requirements to be the work. Turs out the acceptance criteria is what took hours.

Because to write “here is how you know it’s correct,” you have to know exactly what correct looks like. And you have to know what wrong looks like, so you can tell someone how to spot it.

That is the actual skill. Not typing, Proving the config.

It’s also, I think, the thing that separates people who pass the CCNA from people who can do the job. The exam gives you output and asks what it means. A client gives you a broken network and asks the same thing. Configuration is the easy half. Troubleshooting is the half nobody practises.

Writing the acceptance criteria taught me more than building any single lab ever did, and I hadn’t opened Packet Tracer yet.

If you are interested in trying it out feel free to reach out.