Data centre tiers I to IV explained
The Tier levels are Uptime Institute's classification of data centre infrastructure. Each Tier includes the requirements of the ones below it. Tier I is basic capacity, Tier II adds redundant capacity components, Tier III makes the site concurrently maintainable, and Tier IV adds fault tolerance through separate, physically isolated systems. A higher Tier is not a better data centre in itself; Uptime describes the levels as fitting different business operations.
Key facts
- Tier I: basic capacity, including a UPS, dedicated cooling and an engine generator[1]
- Tier II: redundant capacity components for power and cooling[2]
- Tier III: concurrently maintainable, with redundant components and distribution paths[3]
- Tier IV: several separate and physically isolated systems as redundant components and paths[4]
- Higher is not better by default: the levels fit differing business operations[5]
What each Tier means for the UPS
At Tier I, the UPS is there to ride through sags, outages and spikes [1], but there is no redundancy: servicing the UPS means planning an outage or moving the load to bypass.
At Tier II, the UPS capacity has redundant components [2], typically an N+1 module or unit, so a single failure can be absorbed. The distribution path is still single.
At Tier III, everything in the path can be taken out for planned work without shutting down the IT load [3]. For the UPS, that usually means two distribution paths and enough capacity to carry the load while one UPS or one path is out for service.
At Tier IV, the systems are redundant and also physically isolated from one another [4], so a fault or a fire in one cannot take out the other.
The Tiers side by side
| Tier | UPS redundancy | Distribution paths | UPS service without an outage |
|---|---|---|---|
| I | None [1] | One | No: a shutdown or a period on raw utility through the bypass |
| II | Redundant capacity components [2] | One | Module or unit work, yes; work on the shared path, no |
| III | Redundant components [3] | Redundant [3] | Yes, while the redundancy is real |
| IV | Separate, physically isolated systems [4] | Separate and isolated [4] | Yes, and a fault in one system leaves the other running |
Tiers and maintenance
The practical difference most facility managers feel is between Tier II and Tier III. In a Tier II room, some UPS work still needs a shutdown window or a period on raw utility through the bypass. In a Tier III room, the design lets that work happen with the load protected, but only if the redundancy is still real: see how N+1 quietly disappears.
TIA-942 ratings use their own definitions and do not map one to one onto Uptime’s Tiers. Check which scheme a specification means before comparing sites.
Standards cited
- Uptime Institute Tier Classification System (as published)
Common questions
What is a Tier III data centre?
In Uptime Institute's words, a Tier III data centre is concurrently maintainable, with redundant components and redundant distribution paths to the critical environment [3].
Is a Tier IV data centre better than Tier III?
Not automatically. Uptime states that the progression does not make a Tier IV data centre better than a lower one; the levels fit different business operations [5].
Does a Tier I data centre have a UPS?
Yes. Uptime's Tier I requirements include a UPS for power sags, outages and spikes [1].
What does concurrent maintainability mean for UPS work?
That any part of the power path, the UPS included, can be taken out for planned maintenance without shutting down the IT load, because a redundant component or path carries it meanwhile.
Sources
Updated