Four times twenty is sixty

N+1 is one of those labels that stick everywhere and rarely get recalculated. It appears in operations manuals, in contingency plans and in audit responses, and it usually sits on the individual unit.

But redundancy is not a property of a device. It is a property of the room.

The arithmetic the label does not do

A room has four cooling units of 20 kilowatts each, so 80 are installed. The heat load is 65 kilowatts. Each of the four units is properly recorded as N+1.

The room is still not N+1. Because N+1 means it also works when the largest unit fails. If one unit fails, 60 kilowatts remain. The load is 65. Five are missing.

Four times twenty is eighty, but in the redundancy case it is sixty. Exactly that gap between the recorded label and the derived reality is the finding that matters.

Only one direction gets reported. Whoever delivers less than promised gets a warning. Whoever delivers more than promised has no problem and needs no message.

And if it does happen: how long?

The second question follows immediately. If cooling fails, how much time is left before it becomes critical?

That is the most delicate number in the whole module, precisely because it looks so convincing. "Critical in 14 minutes" reads like the result of a flow simulation, and whoever reads it plans accordingly.

A strict order of precedence therefore applies. Where temperature history exists, the actually observed heating rate counts. And not the largest measured hourly rise but the second largest, because the largest may well have been a sensor replacement.

Without history there is a physical approximation via the heat capacity of the room air, with the assumptions spelled out and a visible marker that it is an approximation.

And without a measured temperature there is no time estimate at all. A time starting from an assumed initial temperature would be an invented number, and invented numbers are worse than missing ones in a contingency plan. The starting point is also the warmest spot in the room, not the average.

The same rule one floor down

The same principle applies on the power side. A three-phase system is designed for equal load across all three line conductors. With imbalance, the difference returns via the neutral conductor, and in older installations that is dimensioned weaker than the line conductors.

An installation can sit at the limit on one phase at 60 percent total utilisation, without any total figure showing it. The sum does not reveal the imbalance.

Nothing is invented here either: a circuit whose line conductor is not recorded does not enter the balance and is not tentatively assigned to L1. An invented assignment showed a dramatic imbalance that does not exist, and because that looks like a finding, somebody would act on it.

What it comes down to

Three numbers that matter in operations appear in no datasheet: what remains when the largest unit is missing, how quickly the room heats up, and how unevenly the load sits across the phases.

More on the DCIM module page.