Every data centre design conversation eventually turns to redundancy: how much spare capacity, and where. The shorthand — N, N+1, 2N — is widely used but not always consistently understood. For a facility supplied by onsite generation rather than the grid, it matters even more, because the power plant is the supply.
What "N" means
N is the amount of capacity needed to carry the full design load, with nothing spare. If a facility needs 8 MW and is supplied by generating blocks of 2 MW each, N is four blocks.
A system built to N has no margin: if any one block trips or is taken down for maintenance, the load cannot be fully supplied.
N+1
N+1 adds one extra unit beyond N. In the example above, five 2 MW blocks supply an 8 MW load. Any single block can fail or be maintained without affecting supply.
N+1 is the most common starting point. Its limitation is that it protects against one failure at a time. If a second unit fails while the first is being maintained, supply is at risk.
Variants such as N+2 add further spare units for more protection, particularly where maintenance takes units out of service regularly.
2N
2N duplicates the entire supply path: two fully independent systems, each able to carry the full load on its own. The facility's IT equipment is typically dual-corded, drawing power from both "A" and "B" paths.
2N protects against failures not only in generating units but also in switchgear, transformers, cabling and controls — anything in one path can fail without affecting the other.
2N+1
2N+1 is 2N with an additional spare unit, so the facility can survive a failure in one path while the other path is also carrying a failure or undergoing maintenance. It is used where the cost of any outage is extremely high.
Redundancy is more than spare engines
With onsite generation, it is tempting to count generating blocks and stop. A complete redundancy design also covers:
- Switchgear. A single switchboard is a single point of failure, however many generators feed it.
- Transformers and feeders. Separate A and B paths to the data halls.
- Controls. The microgrid controller and protection systems need redundancy and safe failure modes.
- Fuel supply and storage. Separate feeds and buffer storage so a fuel-side issue cannot stop all units.
- UPS and batteries. Ride-through while standby capacity starts and synchronises.
- People and spares. Operators on site and critical spares on the shelf turn a failure into a short event rather than an outage.
How availability and redundancy relate
Availability is the outcome; redundancy is one of the main tools for achieving it. Moving from N to N+1 to 2N raises achievable availability, with diminishing returns at each step and rising cost. We explain the arithmetic in What 99.999% availability really means.
Choosing the right level
| Question | Pushes towards |
|---|---|
| Is onsite generation the only supply? | More redundancy onsite (at least N+1, often 2N on the distribution side) |
| Is there a reliable grid connection as an extra path? | The grid can act as part of the redundancy |
| What availability do your customer contracts require? | Higher commitments need 2N-style distribution |
| How often are units maintained, and for how long? | Frequent or long maintenance favours N+2 or more |
| How is the facility phased? | Redundancy should be maintained at every phase, not just at full build-out |
How CNP designs for data centres
CNP builds modular onsite generation of 1–20 MW per site for data centres, designed with full backup and redundancy and able to run fully off-grid. Blocks, switchgear and controls are configured to the facility's tier design, and CNP offers contractual availability of up to 99.999% at the point of supply. See Data centre power.
Next step
Assess your site and we will propose a redundancy configuration for your load and phasing.
