Tier levels explained
Tier I to Tier IV is the most famous way to describe how resilient a data centre's design is. Here is what each level really means.
What the Tier system is
The Tier Classification System from the Uptime Institute rates a data centre's power and cooling infrastructure design (and, separately, how the built facility performs) from Tier I (basic) to Tier IV (fault tolerant).
| Tier | Name | What it means in practice |
|---|---|---|
| I | Basic capacity | One path for power and cooling, no spare components. Maintenance means shutting down. |
| II | Redundant capacity components | Spare components (e.g. an extra UPS module or chiller), but still a single distribution path. |
| III | Concurrently maintainable | Any single component or path can be taken out for planned maintenance without shutting down IT. |
| IV | Fault tolerant | The site keeps running through any single unplanned failure, with independent, physically separated systems. |
The two phrases to remember
- Concurrently maintainable (Tier III): you can maintain anything without an outage.
- Fault tolerant (Tier IV): a single failure won't cause an outage either.
Watch out: you'll often see uptime percentages quoted for each tier (like "99.982%"). They are widely repeated, but they are not part of Uptime's official standard — the tiers are defined by design behaviour, not by a percentage.
Design vs. built facility
Uptime certifies the design documents, the constructed facility, and separately how well a site is operated (Management & Operations). A site can be designed to Tier III but run poorly — operations matter as much as design.
Remember
- Tier III = maintain without shutdown. Tier IV = survive a single failure.
- Tiers rate the design and the build, not marketing claims.