Data center tier classification is a way of rating a data center’s power and cooling infrastructure on a scale from Tier I to Tier IV, according to how well the design is meant to tolerate planned maintenance and unexpected equipment failures. The best-known scheme is the Tier Classification System created by the Uptime Institute, which also certifies facilities against it. Other frameworks, such as the TIA-942 standard’s Rated-1 to Rated-4 levels, use similar ideas with their own criteria. Buyers see tier labels most often when comparing colocation providers.
At a glance
- Tiers I to IV describe progressively more resilient power and cooling designs, from basic capacity to fault tolerance.
- The Uptime Institute created the Tier system and describes itself as the only organization licensed to certify against it.
- Certification can apply to a design, a constructed facility or ongoing operations; a provider claiming “Tier III design” may not hold a facility certificate.
- Other frameworks, such as TIA-942 ratings, are similar in concept but are separate schemes with different criteria.
- A tier is a design rating, not an uptime commitment; read the service level agreement for that.
What problem it solves
Every data center claims to be reliable. Without a common vocabulary it is hard to compare one facility’s redundant power and cooling with another’s, or to know whether a site can service its generators without putting your equipment at risk. Tiers give buyers and designers a shared shorthand for how much resilience a data center is built for.
That shorthand helps match facilities to workloads. A test environment or a disposable cache does not need the same infrastructure as a payment platform. Tiers let you ask for a level of resilience, compare quotes on a like-for-like basis and avoid paying for more than you need.
How it works
Each tier adds to the requirements of the one below. The summary below follows the Uptime Institute’s published descriptions in plain terms; the official criteria are more detailed.
| Tier | Common name | What the design is meant to do |
|---|---|---|
| Tier I | Basic capacity | Provides a UPS, dedicated cooling and a generator, but maintenance requires a shutdown and many failures interrupt the IT equipment. |
| Tier II | Redundant capacity components | Adds spare power and cooling components so some can be serviced without a shutdown; the single distribution path remains a weak point. |
| Tier III | Concurrently maintainable | Adds redundant distribution paths, designed so components can be maintained or replaced without shutting down the IT equipment. |
| Tier IV | Fault tolerant | Uses independent, physically isolated systems designed so a single failure or maintenance event does not interrupt the IT equipment. |
Redundancy terms. Tier designs are often described using redundancy shorthand such as N+1 (one spare component beyond what the load needs) or 2N (two complete systems). Redundancy alone does not determine the tier; the distribution paths and how the systems are separated matter as well.
Certification. The Uptime Institute offers certification at different stages, covering the design documents, the constructed facility and the facility’s operations. Its website lists certified facilities. A facility can be designed to a tier without being certified, and a design certificate does not by itself show that the building was constructed as designed.
Other frameworks. The Telecommunications Industry Association’s TIA-942 standard defines Rated-1 to Rated-4 levels and covers telecommunications, architectural and other areas as well as power and cooling. Other standards, such as the European EN 50600 series, use their own classes. These are separate from Uptime’s Tiers, with separate assessors.
For help comparing facilities and their ratings, see our Colocation solution page.
When it matters for buyers
- Choosing a colocation site. Ask which tier the facility is designed to, who rated it and what was certified.
- Placing critical systems. Systems that cannot tolerate downtime may justify a higher tier, or a second site, or both.
- Comparing prices. Higher tiers cost more to build and run; make sure you are comparing like with like.
- Customer and audit requirements. Some contracts or customers specify a minimum tier or rating for hosting facilities.
Questions to ask vendors
- Which tier is the facility designed to, and which scheme are you using (Uptime Institute, TIA-942 or another)?
- Is the rating certified, and for which stage: design documents, constructed facility or operations? Can we see the certificate?
- If not certified, what evidence supports the claim?
- How are power and cooling maintained, and has any maintenance required shutting down customer equipment?
- What outages has the facility had in recent years, and what caused them?
- How does the service level agreement for power and cooling relate to the tier claim?
How it differs from a service level agreement (SLA)
A tier describes how the infrastructure is designed (and, if certified, built or operated) to handle maintenance and failures. A service level agreement is the provider’s contractual commitment, such as power availability to your cabinet, with credits if it misses. A Tier III facility can have a weak SLA and a facility with no tier claim can offer a strong one, so evaluate both: the tier tells you how likely failure is by design, and the SLA tells you what the provider owes you when it happens.