Audits and planning
A redundant design stays redundant only while the equipment left after one failure can carry the whole load. Loads grow, servers are added to one side, modules are taken out and not replaced, and a design drawn as N+1 or 2N can quietly become N. A capacity and redundancy review measures the load on each UPS module and each power path and compares it with what would remain after the loss of one unit.
A count of what is needed and what is spare. Uptime Institute describes N as the number of components minimally required to meet the load, with more equipment described as N+1, N+2, 2N or 2(N+1). Vertiv describes 2N as two wholly independent, parallel UPS systems, each able to support the entire load on its own. The definitions are explained in UPS redundancy; this review asks a different question: whether the site, as loaded today, still behaves the way the design says.
Through load growth and drift. A 2N pair keeps its redundancy only while either side can carry the whole load alone; once the total load is more than one side can carry, losing a side drops part of the room. An N+1 modular UPS keeps its redundancy only while the load fits on the modules that would remain after one fails. Vertiv's paper describes what happens at the edge: to meet the total rating after a module is lost, the remaining modules run in overload. Nothing alarms while all units are healthy, so the loss of redundancy is invisible until the day it is needed. ABB's reliability figures show what is at stake: for a four-module modular UPS, it calculates an MTBF of 1,248,000 hours with a redundant module and 111,111 hours without one.
Each figure is compared with what would remain after the loss of one unit. The answer for each system is a plain yes or no: does the redundancy hold at today's load, and how much load can be added before it does not?
Because an unbalanced pair loses redundancy sooner. Uptime Institute's description of dual-corded equipment expects the normal load on each power source to be shared within 10% of the average when both inputs are available, and on three-phase sources the load on each phase within 10% of the average as well. When a review finds one side carrying far more than the other, the usual causes are single-corded equipment on side A, servers with both cords on the same side, or a failed power supply that has pushed its load to the other cord. Each is a finding, because each shrinks the margin the design was built around.
Capacity that is installed and paid for but cannot be used. Vertiv warns that when loads are not balanced across circuits and phases, operators risk overloading circuits or stranding power: one phase or circuit runs near its limit while others sit lightly loaded, and the full rating of the UPS or PDU cannot be reached without tripping something. Stranded capacity also comes from redundancy itself: in a 2N design, at most the rating of one side is usable, however much is installed. A review separates the two, so the owner knows how much real headroom exists before buying more.
| Finding | Typical fix |
|---|---|
| Load above what one side of 2N can carry | Move load, add capacity, or accept and document the lost redundancy |
| N+1 modular UPS without spare capacity | Add a module, or reduce the load on that UPS |
| Unbalanced A and B sides | Re-cord servers and move single-corded loads to transfer switches |
| Unbalanced phases or circuits | Rebalance across circuits and phases at the PDU |
| Runtime short at today's load | Battery replacement or extension, sized to the measured load |
Battery runtime is part of the same review, because a UPS that carries the load but cannot hold it until the generator is online is not doing its job; batteries for data-centre UPS covers formats and sizes.
Whenever the load changes materially, and at least with each maintenance cycle. The measured loads come from the UPS display, the floor PDUs and monitored rack PDUs, so a site with monitored rack PDUs can track it continuously. The review is also the starting point for any expansion: before adding a UPS module or a new row of racks, it shows where the capacity is and which redundancy it would affect.
Yes. Send the UPS and PDU list, the single-line diagram and any recent load readings through the audit quote request. The review is often combined with data-centre UPS maintenance, since the readings are taken at the same equipment.
N+1 adds one spare unit to the N needed; 2N, as Vertiv describes it, is two wholly independent systems, each able to support the entire load.
Yes. If the total load grows beyond what one side can carry alone, the loss of one side drops part of the load, even though both sides are healthy.
Installed capacity that cannot be used, often because loads are unbalanced across circuits and phases, or because a redundant design only allows one side's rating to be used.
For a four-module modular UPS, ABB calculates an MTBF of 1,248,000 hours with a redundant module and 111,111 hours without.
Sources, checked 2026-10-04:
Tell us the site, the UPS and the timing. We reply and quote within 4 business hours, or call 1 (438) 800-0382
Brand names are trademarks of their owners and are used only to identify the equipment we service. We are not the manufacturer and are not affiliated with it.