Power
Data Centre Power Explained: kW, Feeds and Redundancy
7 min read
Understand kilowatts, electrical feeds, PDUs and redundancy before comparing data centre colocation proposals.
What a kilowatt means in colocation
A kilowatt, or kW, measures the rate at which equipment uses electrical power. In a colocation proposal, the kW figure normally describes the IT load available to a rack or customer footprint, but definitions can vary. It is not the same as kilowatt-hours, which measure energy consumed over time and appear on electricity bills. A server may have a high nameplate rating but draw less during normal operation. Buyers should measure actual load, model peak demand and add controlled headroom rather than simply adding every power-supply rating together.
A feeds, B feeds and dual-corded equipment
Many data centres provide two independent power feeds, commonly called A and B. Dual-corded servers can connect one power supply to each feed so they remain online if one path is unavailable. Single-corded equipment may require an automatic transfer switch or another carefully designed solution. Two sockets do not guarantee true resilience: the feeds should be traced through rack PDUs, distribution equipment, UPS systems and generators. Ask the provider to explain where the paths are separate and whether maintenance or a single fault could affect both.
UPS systems and generators
Utility power can fail or fall outside acceptable voltage and frequency ranges. Uninterruptible power supply systems support the IT load immediately, while standby generators start and stabilise. The UPS may use batteries, flywheels or another energy-storage design. Generators then sustain the facility for longer events, subject to fuel arrangements and operating procedures. Redundancy can be expressed as N, N+1, 2N or other configurations. These labels describe capacity design, but buyers should also consider distribution paths, physical separation, maintenance practices and testing.
Usable power and rack density
A provider may offer a nominal power allocation, but the usable amount can depend on feed balancing, breaker ratings, redundancy commitments and cooling capability. For example, a customer wanting resilient dual feeds may be unable to load both circuits to their full rating because either feed must be capable of carrying the environment after a failure. Rack density also affects airflow and equipment placement. Traditional enterprise racks may operate at modest density, while GPU systems can require tens of kilowatts or more per rack. Confirm both present and future density support.
Questions to ask in a power review
Ask whether charges are based on reserved or consumed power, how usage is metered, and what occurs if the agreed limit is exceeded. Request the supported plug types, voltage, breaker size and rack PDU options. Confirm the redundancy design from utility entry to rack outlet and whether the service has concurrent-maintenance or fault-tolerance characteristics. Review generator runtime assumptions, fuel contracts, load testing and incident escalation. Finally, obtain written confirmation that your planned equipment and future growth can be accommodated in the selected hall.
Frequently asked questions
Not sure how this applies to your business?
The free assessment turns answers like these into realistic pricing and a provider shortlist for your specific requirements.
Start free assessment