Tiers are a point of significant confusion in the data centre industry. The origin of Tiers came from the Uptime Institute [founded in 1993 as Site Uptime Network] to help define “uptime” [not redundancy].
The Uptime Institute classified data centres into Tiers in Roman numerals I through IV where Tier IV represents the highest level of projected availability. Their Tiers defined the meter stick for measuring data centre availability. Their original “Tier Standard” summarized the tiers into measured percentage of availability as follows:
- Tier I: 99.671%
- Tier II: 99.749%
- Tier III: 99.982%
- Tier IV: 99.995%
Many years have passed since the first Tier Standard document was written by the Uptime Institute and the more recent release of the document describes the Tiers much better than before.
In 2005, the TIA approved the Telecommunication Infrastructure Standard for Data Centres or TIA-942 Standard. This standard also defines data centres into Tiers. They use Hindu-Arabic numerals, again Tier 1 through 4 where Tier 4 that are based on the Uptime Institute’s Tiers.
The TIA-942 standard is written as guidelines or common/best practices. TIA Tiers are better understood as general levels of redundancy [as opposed to availability with the Uptime Institute].
- Tier 1: N or “just enough”
- Tier 2: N+1 [but not for all systems as generator is still N]
- Tier 3: 2N [dual utility feeds (active and dormant) and N+1 power components]
- Tier 4: 2(N+1) [dual active utility feeds]
For more information on Uptime Institute’s Tiers and TIA-942 Tiers, please visit their respective websites.