Designing a data-center capacity calculator: turning infrastructure complexity into decisions
Chapter 1 of 4 · Designing a data-center capacity calculator
A capacity calculator needs to help someone understand what a facility could support, what it currently uses, and which assumptions connect the two. A single number cannot explain all three.
The PhysaFlow calculator implementation separates these questions. That separation is the most useful place to begin its design story: the meaning of the result determines how someone can use it.
Start with the limiting constraint
The model considers three supplied limits: power available for IT after reliability headroom and PUE, commissioned electrical IT capacity, and cooling-supported IT capacity. It takes the smallest available value as usable IT capacity.
In the illustrative example, gross deliverable power is 100 MW. Subtracting 10 MW of intentional headroom and dividing by a PUE of 1.5 produces a 60 MW power-derived IT limit. Commissioned electrical capacity is 80 MW and cooling-supported capacity is 70 MW. The limiting supplied constraint is therefore 60 MW.
This gives the result an explanation a reader can follow. It also makes the boundary explicit: only supplied constraints participate. Missing constraints are disclosed, as the next chapter explains.
Give each capacity number a distinct meaning
With an average measured IT load of 36 MW, the example has 24 MW of unused IT capacity. The model then applies its air-cooling scenario coefficient of 0.485 to that unused capacity, producing 11.64 MW of modeled stranded capacity.
These are different statements. Unused capacity describes the gap between the modeled usable limit and measured load. Modeled stranded capacity applies an additional assumption to that gap. Potentially recoverable capacity requires a further user-supplied recovery share.
The coefficient is a model assumption; the provided files do not establish its empirical provenance. The resulting 11.64 MW is an illustrative estimate, not a measured finding or a verified recovery forecast.
Design around the next question
The useful output is a sequence of questions: Which supplied constraint limits capacity? How much is unused? What portion does the scenario classify as stranded? What evidence would support a recovery estimate?
For a product team, this creates a clear content hierarchy: explain the physical quantities before attaching a financial value. A dollar figure becomes more useful when readers can trace it back to the capacity definition underneath it.
That is the central design lesson of this implementation: precise labels and visible dependencies help turn a calculation into a basis for further investigation.
This chapter is part of the PhysaFlow calculator series overview. The PhysaFlow case study explains the surrounding positioning and product work. Continue the reasoning with designing useful partial answers when measurements are missing.