An uninterruptible power supply (UPS) is a piece of electrical equipment that keeps connected devices running when utility power fails, dips or becomes unstable. It stores energy, usually in batteries, and delivers it the moment the incoming supply falls outside acceptable limits. UPS units range from small desktop and rack-mounted models protecting a few devices to large systems feeding an entire data center. In most facilities a UPS is the bridge that carries the load until a generator starts or power returns, not a long-term power source on its own.
At a glance
- Takes over from utility power with little or no interruption, using stored energy such as batteries.
- Many models also filter out surges, sags and other power-quality problems during normal operation.
- Runtime is limited, so it is normally paired with generators in data centers or with a clean shutdown plan in offices.
- Batteries age and fail, so testing and maintenance determine whether the protection is real.
- Redundant UPS designs help avoid a single power component taking down the whole load.
What problem it solves
Servers, network equipment and phone systems are sensitive to power problems. A full outage is the obvious one, but short sags, surges and fluctuations can also cause reboots, corrupted data or hardware damage. Even in a building with a generator, there is a gap of seconds between the utility failing and the generator taking the load, and without something to bridge it, equipment simply loses power.
A UPS fills that gap and smooths the supply. In a data center, it keeps the IT load running while generators start and stabilize. In an office or comms room, it rides through brief outages and gives systems time to shut down cleanly if the outage lasts longer. For phones and network equipment, it can keep communications up during short outages, which matters for safety as well as business continuity.
How it works
Stored energy. Most UPS systems use batteries, commonly valve-regulated lead-acid or, increasingly, lithium-ion. Some large systems use flywheels instead. The amount of stored energy and the size of the load determine runtime.
Designs. Three common designs differ in how they handle normal power. Standby (offline) units pass utility power through and switch to battery when it fails. Line-interactive units add voltage regulation, correcting sags and surges without draining the battery. Double-conversion (online) units convert incoming power to direct current and back again continuously, so in normal operation the equipment runs from the UPS’s own output; this gives the cleanest power and no switch-over when the utility fails, at the cost of some efficiency. Data centers typically use double-conversion systems, sometimes with an energy-saving mode.
Redundancy. Facilities arrange UPS modules in redundant configurations, such as N+1, where one extra module can fail without dropping the load, or in two fully separate power paths feeding each cabinet. The configuration is one of the factors in a facility’s tier classification. Downstream, power distribution units (PDUs) carry UPS-backed power to the racks.
Maintenance. A UPS needs periodic battery testing, replacement and servicing. Many have a bypass that lets technicians work on them while the load runs on utility power, which temporarily removes protection unless another UPS path covers the load.
For on-site support of UPS units and other equipment at your locations, see our Field Support solution page.
When it matters for buyers
- Choosing a colocation facility. The UPS design and redundancy determine whether a single equipment failure or maintenance window can interrupt your power.
- Protecting offices and comms rooms. Network switches, firewalls and phone equipment need UPS protection sized to the load and the runtime you expect.
- Replacing copper phone lines. Digital replacements for traditional lines often need local backup power to keep working during an outage; requirements vary by jurisdiction and use, so check with the provider and the authority having jurisdiction.
- After a power-related outage. A UPS with aged batteries or an overloaded output can fail when it is needed, so review testing records and load.
Questions to ask vendors
- What UPS design and redundancy level feeds our cabinets, and are our two power feeds backed by separate UPS systems?
- How long can the UPS carry the load, and how long do the generators take to pick it up?
- How often are batteries tested and replaced, and can we see the records?
- What happens to our power during UPS maintenance or bypass?
- For office equipment: what runtime will this unit give at our actual load, and who monitors battery health and alerts us?
How it differs from high availability
A UPS is a piece of equipment that protects against one cause of downtime: problems with incoming power. High availability is a design approach for keeping a whole service running through many kinds of failure, using redundant components, failover and the removal of any single point of failure. A UPS, or a redundant pair of them, is often one part of a high-availability design, but on its own it does nothing about failed servers, software or network links.