Crash course: data centre tiers

By Harry Menear
Data centre operators boast about being Tier III or Tier IV certified. But what does that mean...

For more than 25 years, the Uptime Institute’s data centre tiering system has been the defining certification in the industry. But what does it mean exactly to operate a Tier III data centre compared to a Tier II facility? 

The Uptime Institute was founded in 1993, and since then has operated as the third-party adjudicator of data centre resilience and complexity. Its four tier certification standards evaluate the quality and reliability of a data centre’s server hosting ability using a proprietary rating system that assesses power, cooling, backup capabilities, redundancies and average outage time per year. It’s also worth noting that the tiering system is progressive, so a Tier III data centre must meet the criteria for tiers I and II, in addition to those required for Tier III. 

Tier I

The most basic data centres. Tier I facilities are essentially warehouses with electricity and cooling. Typically used by smaller businesses and legacy operators, Tier I data centres provide the basic data centre infrastructure that goes one step beyond a server in an office. They have a dedicated space for servers and other IT components, an uninterruptible power supply, dedicated cooling systems that continue operating outside business hours, and a backup generator. Tier I facilities must be able to guarantee 99.671% uptime, and no more than 28.8 hours of downtime per year. 

Tier II

This is where modern data centre design methodology starts to take effect. The average enterprise data centre, or facility in a developing market will have capabilities equal to a Tier II certification. Tier II facilities are defined by their redundant capacity, including critical power and cooling redundancies that establish “an increased margin of safety against IT process disruptions that would result from site infrastructure equipment failures,” according to Matt Stansberry of the Uptime Institute. 

Tier III

The highest level of certification you’ll ever need unless you’re running a public cloud, cutting edge research facility or maybe a government. Tier III data centres are concurrently maintainable, which means they require no disruption of service even when conducting major repairs, replacements or maintenance. Also, the certification requires additional redundancy of power and cooling, “so that each and every component needed to support the IT processing environment can be shut down and maintained without impact on the IT operation.” 

Tier IV

So, you’ve got a major public cloud, average-sized government or respectable military to run and you need to put your data somewhere. The redundancies and capabilities provided by Tiers I through III are further built upon by Tier IV data centres with the addition of dramatically increased fault tolerance measures. “Fault Tolerance means that when individual equipment failures or distribution path interruptions occur, the effects of the events are stopped short of the IT operations,” writes Stansberry. 

In conclusion, he reflects that the “Uptime Institute recognises that most data center designs are custom endeavors, with complex design elements and multiple technology choices. With all of the technology choices available to designers today, including renewable power sources, innovative distribution and advanced cooling approaches, the Tier Standard of Topology is unique in being able to encourage that usage, as long as the resulting infrastructure precisely meets the outcomes defined in the Tier Standard for the level chosen. As such, the Tier Classification System does not prescribe any specific technology, schematic or other design criteria beyond those resulting outcomes stated above. It is up to the data center owner to meet those outcome criteria in any method that fits his or her infrastructure goals based on the business parameters.” 

Whatever your business demands, the Uptime Institute’s certification tiers are a solid guide to understanding the data centre that best suits your needs. 


Featured Articles

US academics warn of huge AI energy requirements

AI’s potential pitfalls are becoming clear, with worries growing over job losses and privacy concerns; the world can add environmental damage to the list

5 minutes with: Zachary Smith, Equinix

In the latest edition of Data Centre Magazine, we hear from Zachary Smith, who runs the global edge infrastructure services of data centre leader, Equinix

Peter Herweck has been named as Schneider Electric’s new CEO

Peter Herweck, the former CEO of AVEVA, will replace Jean-Pascal Tricoire as the CEO of sustainable data centre giant Schneider Electric

Svein Atle Hagaseth appointed as Green Mountain’s new CEO

Data Centres

Top 10 data centre marketing executives


On the pod – unexpected medium for data centre marketing