A practice running on individual engineers’ equipment is shadow IT with a job title. That is the whole argument, and it is the reason this was worth doing before anything more interesting was.
What it is
Three servers already owned and no longer doing what they were bought for, being rebuilt into a single governed virtualized network emulation cluster carrying 2.3 TB. Not new capital — recovered capital, which is what made the case winnable.
Why it had to be argued rather than requisitioned
Standard procurement lead time was six months. A practice with twenty-plus concurrent engagements does not have six months to wait for the environment in which it validates a customer design, so the alternative was the status quo: architects testing on whatever they personally owned, with no shared state, no governance and nothing anybody else could reproduce.
The case was made on cost per month against that lead time — the recovered hardware set against the delay, rather than against a competing purchase. Framing a funding argument as this versus waiting rather than this versus that is usually the difference between an approval and a queue position.
What it carries
Three uses at once, which is the other half of the justification:
- The network orchestration proof of concept, where Itential and Gluware were built out on virtualized topologies rather than judged on datasheets.
- The thirty-six-month post-quantum cryptography program with the group’s parent company.
- Pre-production design validation for client work — customer designs built and tested in the lab before anything touched a production network.
That last one is the point of the whole exercise. A design review can catch what a drawing says; only a build catches what the drawing left out.
Outcome
Customer designs built and tested before touching production, on equipment the company owns rather than equipment individual engineers do.