01 — The situation
Eight sites, eight control rooms, no shared screen.
Our client handles instrumentation across eight power plants. Each site has its own measuring equipment, and the readings were already being collected there — but could only be seen from that site's control room.
What was needed: monitoring screens that could be opened from outside the control room, looking the same at every site so nobody has to relearn them when they move.
02 — The problem
Alike, but not identical.
Eight sites sounds like one job multiplied by eight. It isn't.
Each site has equipment of different makes and ages, and the way readings are collected differs accordingly. What is identical is only the shape of the question: what is the value now, what has the trend been over the last few hours, and is anything past its safe limit.
The great temptation in work like this is to build one enormous system serving all eight. That means one troubled site can take down the screens of the other seven.
03 — What we did
One pattern, eight installations that stand alone.
Four decisions.
- 01
The collector separated from the display
The only thing that differs per site is how readings are collected. That part is a separate bridge, while the display is identical everywhere. Adding a new site means writing its bridge, not rebuilding the dashboard.
- 02
Each site runs on its own, not on a shared central system
Eight separate installations. If one site has trouble — its network drops, its equipment stops — the other seven are entirely unaffected.
- 03
Raw readings stored untouched, summaries computed on top
Field values are stored exactly as received. Averages and trends are a layer above them. If the way a summary is computed ever needs to change, the original data is still intact and can simply be recomputed.
- 04
Built to be read at a glance, not studied
Monitoring screens are often watched from a distance and for seconds at a time. Anything outside its safe limit must stand out without being read item by item.
04 — The outcome
Eight sites installed within two years.
All eight use the same dashboard pattern, each with its own bridge to the equipment on the ground. The later sites went in considerably faster than the first — because all that remained to write was the bridge.
05 — What keeps it standing
What gets reused isn't the code. It's the shape of the problem.
Copying code from one site to the next produces eight copies that slowly drift apart until nobody dares touch any of them.
What we reuse is the separation — readings in, stored raw, summarised, then displayed. That pattern is why a ninth site, if there is one, would be no harder than the eighth.
