We adopted a cloud-first strategy years ago and it was the right call. What nobody wrote down is that cloud-first was quietly read as cloud-only, and roughly forty percent of our workloads are still sitting in two comms rooms and a leased rack in a facility twenty minutes up the road.
The problem is not the remaining hardware. It is that every scrap of planning attention, training budget and architectural interest moved to the other side of the estate, and the physical half kept running the company without anyone thinking about it. We found out how far it had drifted over one weekend when an air conditioning unit failed, the temperature alarm went to a mailbox that had been decommissioned in a tidy-up, and the only person who knew which rack held what had retired eighteen months earlier. Two hardware support contracts had lapsed. The uninterruptible power supply batteries were four years past their replacement date. There was no current diagram, only a spreadsheet from a migration that had been abandoned halfway.
What sharpened it for me is which workloads stay behind. They are never the trivial ones. They are the systems tied to a licence bound to a physical machine, the appliance with no cloud equivalent, the application whose vendor has eleven customers and no roadmap, the thing a regulator has opinions about. The residue of a migration is by definition the least movable part of the estate, which usually means the most entangled and often means the most important.
So we stopped calling it legacy and gave it an owner, a refresh plan with actual dates, and a line in the budget that recurs. Environmental monitoring that pages a rota rather than a mailbox. Hardware support renewed on the things that matter, with the risk on the rest written down and accepted by a person rather than by silence. A walk-through once a year with photographs. At least two people who can physically get in.
Hybrid is not a stage we are passing through. For most organisations I know it is the steady state. Calling something legacy describes where our attention went, not how much of the business depends on it.
– Serguey Shinder
Top comments (0)