What is redundancy?
Having a spare, so that when one part fails the service carries on.
Duplicate what cannot be allowed to fail.
Redundancy means adding extra capacity, components or routes so that a failure in one does not stop the whole service. A spare power supply, a second network route, a backup generator and a second data centre are all examples. Redundancy is one of the main ways to remove a single point of failure.
It costs money, adds complexity and must be tested. Used well, it buys time and reduces the chance that one fault becomes an incident.
N, N+1 and 2N explained.
| Level | Meaning | Example | Trade-off |
|---|---|---|---|
| N | Just enough capacity for the load. | One chiller sized for the whole data hall. | No spare. Any failure or maintenance affects service. |
| N+1 | The needed capacity plus one spare. | Four chillers where three are needed. | Survives one failure. Cost is moderate. |
| N+M | Several spare units. | Six generators where four are needed. | Survives several failures, but larger and costlier. |
| 2N | A fully duplicated system. | Two independent power paths, each able to carry the full load. | Survives the loss of a whole path. High cost. |
| 2N+1 | Duplicated, plus a spare. | Two full systems, with an extra unit. | Highest protection, highest cost. |
Names and meanings can vary between organisations and standards. Check the definition used in your contract or design.
Where to apply it.
Power
Two utility feeds, UPS, generators, dual power supplies in equipment and separate distribution paths.
Cooling (HVAC)
Spare chillers, pumps and units and separate loops. See What is HVAC?.
Network and telecoms
More than one carrier and route, diverse building entries, redundant routers and links.
Data and systems
Replication, clustering, active-active or active-passive design and backups. See data centre architectures.
Sites and locations
A second site or work area recovery site, far enough from the first to avoid shared risks. See What are Work Area Recovery Sites?.
Suppliers
More than one source for critical goods and services, or approved alternatives.
People and skills
Deputies, cross-training and succession for key roles, so knowledge does not sit with one person.
Processes and manual workarounds
Paper or manual methods that can run when systems are down.
Why redundancy sometimes fails.
Common-cause failure
Two systems that share a hidden dependency, such as the same power feed, cable duct, software or supplier, can fail together.
Untested failover
A spare that has never been tested may not start, or may not take the load.
Silent failure
A backup that failed unnoticed leaves you with no protection. Monitor spares as well as live systems.
Complexity
More components mean more that can go wrong, and failover logic can itself cause outages.
Capacity creep
The load grows until the spare is needed just to run normally.
Human error
Mistakes during maintenance or change often cause the outage that redundancy was meant to prevent.
Redundancy lessons from real events.
March 2025, Heathrow substation fire
A fire at an electrical substation was reported to have cut power to the airport, which closed for most of a day, raising questions about backup supplies and shared infrastructure.
October 2021, Facebook outage
A configuration change cut data centres off from the internet, and internal tools needed for recovery were affected too. Redundant systems shared the same dependency.
October 2012, Hurricane Sandy
Flooding disabled fuel pumps and generators in lower Manhattan, showing that backup power depends on fuel supply and location.
April 2025, Iberian blackout
A large grid failure reminded organisations that generators, batteries and manual processes need to be ready, tested and sized for the load.
Examples are drawn from public reports and are simplified. Check the original reports for details.
Redundancy questions.
Is redundancy the same as backup?
How much redundancy do we need?
Can redundancy be too much?
Talk to ADVOX about resilience design.
We aim to reply to your enquiry within 2 hours during working hours.
Related reading: What is Redundancy? | What is a Single Point of Failure? | What is Concentration Risk? | What are Data Centres? | What is HVAC?