Resources » Data Center Power Terms Explained: A Plain-English Glossary

Data Center Power Terms Explained: A Plain-English Glossary

by | Oct 5, 2026

Data center power terminology mixes electrical engineering, IT operations and procurement vocabulary, which is why the same equipment gets three names depending on who is talking. This glossary covers the terms that come up most in real projects: capacity terms like kVA, kW and power factor; the distribution chain from switchgear down to the rack; redundancy notation such as N+1 and 2N; and the monitoring levels that separate a basic power strip from an intelligent PDU.

Capacity and load terms

These are the terms that cause the most confusion, because two of them look interchangeable and are not.

Capacity and load
Term What it means Why it matters
kVA (apparent power) Volts times amps, divided by 1,000. What the electrical infrastructure has to carry. Equipment and breakers are sized in kVA. Ignoring the distinction undersizes infrastructure.
kW (real power) The power actually converted into work and heat. Cooling and utility billing follow kW. IT equipment is usually rated in kW or watts.
Power factor (PF) kW divided by kVA. How efficiently apparent power becomes real power. Modern IT power supplies are typically near 0.99, so data center designers often treat kW and kVA as close to equal. Older or mixed loads may not be.
Single phase Two conductors plus ground. One alternating voltage. Common for smaller loads and legacy racks.
Three phase Three conductors plus neutral and ground, offset in time. Delivers more power over less copper. Standard for higher-density racks.
Nameplate rating The maximum draw printed on equipment. Almost always higher than real draw. Designing to nameplate strands capacity.
Actual draw What equipment really pulls under its normal workload. The number worth designing around, once you can measure it.
Diversity The gap between the sum of nameplate ratings and actual simultaneous draw. Not everything peaks at once. Diversity is what lets a facility support more nameplate than its feeder rating.
Derating Reducing usable capacity below the rated maximum for safety and code compliance. The most commonly forgotten step in rack sizing.
Continuous load A load expected to run at maximum for three hours or more. Electrical codes generally require continuous loads to be limited to 80 percent of the circuit rating. Almost every IT load qualifies.
Stranded capacity Power that exists somewhere in the chain but cannot reach the equipment that needs it. Usually caused by running out of breaker positions, not out of amps.

The distribution chain

Power moves through a sequence of equipment between the utility and the server. These are the pieces, roughly in order.

Distribution equipment
Term What it means Where it sits
Transformer Changes voltage between one level and another. Between the utility and the facility, and again inside the building
Switchgear Metal-clad or metal-enclosed assembly with draw-out breakers, built for higher fault levels. Primary protection and switching, upstream
Switchboard Fixed-mounted assembly used for distribution rather than primary switching. Downstream of switchgear, feeding the facility
UPS Uninterruptible power supply. Rides through disturbances and bridges to generator. Between the building supply and the critical load
PDU (facility) Floor-standing unit that transforms and distributes power to the data hall. Room level
RPP Remote power panel. A distribution panel placed out on the floor, closer to the load. Row or zone level
Busway Enclosed conductor rail, usually overhead, that power taps off along its length. Above rack rows
Tap-off box The unit that draws power off a busway run and feeds equipment. Contains its own protection. Attached to the busway
Power whip Pre-terminated cable assembly running from a panel to a single rack. Panel to rack
Rack PDU The power strip inside the rack that servers plug into. Also called a CDU or rPDU. Inside the rack
Receptacle / outlet The socket itself. Type determines voltage and current. On the rack PDU or tap-off box

The word PDU causes most of the confusion in this list. A facility PDU is a floor-standing transformer and distribution cabinet. A rack PDU is a power strip. They are separated by two orders of magnitude in capacity and cost, and both get called ‘the PDU’ in the same meeting. When it is ambiguous, say ‘rack PDU’ or ‘floor PDU’ explicitly.

Redundancy notation

The N notation describes how much spare capacity exists and how it is arranged.

Redundancy terms
Term What it means Practical effect
N Exactly the capacity needed to carry the load. No spare. Any failure drops load.
N+1 One redundant component beyond what the load requires. Survives a single component failure. Shares a common path.
2N Two complete independent systems, each able to carry the full load. Survives a failure and allows maintenance on one side while the other runs.
2N+1 Two complete systems, one carrying a spare component. Redundancy that survives a failure during maintenance.
A/B feed Two independent power paths, conventionally labelled A and B. The physical implementation of 2N at the rack.
Dual-corded Equipment with two power supplies, each fed from a different path. Lets a rack survive losing one entire side.
Single-corded Equipment with one power supply. Needs a static transfer switch to benefit from A/B feeds.
STS Static transfer switch. Switches a single-corded load between two sources. How single-corded equipment gets redundancy.
Concurrent maintainability The ability to service any component without shutting the load down. The practical reason most facilities specify 2N distribution.

Monitoring and control levels

Rack PDUs are sold in tiers, and the names are used loosely across manufacturers. These are the functional distinctions.

Rack PDU capability levels
Term What it does What it does not do
Basic Distributes power. Nothing else. No measurement, no network connection
Metered (unit level) Displays total current for the whole strip, usually on a local display. Cannot tell you which outlet is drawing what
Monitored (networked) Reports unit-level measurements over the network to a management system. Still no per-outlet visibility
Outlet-metered Measures current at each individual outlet. Cannot switch outlets on or off
Switched Turns individual outlets on and off remotely. Not all switched units also meter per outlet
Switched and outlet-metered Both per-outlet measurement and per-outlet control. The most expensive tier
Branch circuit monitoring Measures current per branch circuit rather than per outlet. Coarser than outlet metering, often sufficient

A useful buying check: ‘monitored’ and ‘metered’ are used inconsistently between manufacturers, and the difference between unit-level and outlet-level measurement is a large price step. Ask what is measured and at what point, rather than relying on the product tier name.

Terms worth getting right in a specification

A few terms are frequently used loosely in conversation and precisely in a specification. Getting them wrong produces equipment that arrives correct and useless.

  • Voltage notation. 208/120V describes a three-phase four-wire system where phase-to-phase is 208V and phase-to-neutral is 120V. Writing ‘208V’ alone leaves the neutral arrangement ambiguous.
  • Receptacle type. NEMA designations such as L6-30 or L21-30 encode voltage, current and configuration. The letter L means locking. Getting this wrong is the single most common cause of equipment arriving that physically cannot connect.
  • Amps versus kW. Circuits are rated in amps; IT loads are described in watts or kW. Converting between them requires the voltage and, for three phase, a factor of the square root of three.
  • Phase balance. On a three-phase circuit, loads should be distributed roughly evenly across phases. A rack PDU that is fully loaded on one phase and empty on another wastes capacity and can trip protection early.

Frequently asked questions

What is the difference between kVA and kW?

kVA is apparent power, calculated from volts and amps, and represents what the electrical infrastructure must carry. kW is real power, the portion actually converted into work and heat. The ratio between them is power factor. Modern IT power supplies typically run at a power factor near 0.99, so the two figures are close in most data center contexts, but equipment and breakers are sized in kVA and cooling is sized in kW.

What is the difference between a PDU and an RPP?

A facility PDU is usually a floor-standing unit that transforms voltage and distributes power to the data hall. A remote power panel is a distribution panel placed out on the floor, closer to the racks, typically fed from a PDU or switchboard and without its own transformer. Confusingly, a rack PDU is a different thing again: the power strip inside the rack. Saying ‘floor PDU’ or ‘rack PDU’ avoids most of the ambiguity.

What does dual-corded mean?

Dual-corded equipment has two power supplies, each connected to a different power path, conventionally called A and B. If one path fails or is taken down for maintenance, the other carries the equipment. It is the mechanism that makes 2N redundancy meaningful at the rack. Single-corded equipment needs a static transfer switch to get similar protection.

What is stranded capacity in a data center?

Stranded capacity is power that exists somewhere in the distribution chain but cannot reach the equipment that needs it. The most common cause is running out of breaker positions or outlets rather than running out of amps: a panel can be physically full while its feeder still has headroom. It is a design and layout problem rather than a capacity problem, which is why adding more upstream capacity does not fix it.

Next step

Trilogy engineers and supplies data center power distribution, from switchboards and transformers through busway and rack-level distribution. If a specification is throwing up terms you want a second opinion on, we are happy to help.

Related Articles

Ready To Get Started?

Tell us about your next project!
Speak with a Trilogy team member today.