Data centers must be able to meet their clients’ needs successfully. Because businesses and their needs vary, a data center that can accommodate one type of business client may not be suited to take on another kind, and vice versa.
The Uptime Institute, as a division of 451 Group, developed a system to certify data centers based on the demands the center is able to reliably meet. These data center tier ratings have been in use since the mid-1990s, and they enable business clients to determine if a given data center will be capable of handling their needs.
What Are The Data Center Tiers?
Data center tier certifications fall into one of four groups, classified by Roman numerals I through IV. If a data center advertises their tier in a non-Roman numeral format, it may mean that the certification did not come from the recognized Uptime Institute and the certification might have entirely different criteria for meeting the particular level.
Official Uptime Institute certifications are awarded based on a facility’s design documents, construction, and operations.
- A Tier Certification of Design Documents measures a facility’s design and plans against criteria set by the Institute and it proves that the data center’s design meets those standards.
- The Tier certification of Constructed Facility measures the finished, constructed capabilities against the design documents and further criteria from the Institute.
- The Tier Certification of Operational Sustainability (TCOS) shows that a data center is able to maintain a suitable level of efficiency and availability in daily operations.
Most businesses looking for a data center should be primarily concerned with the TCOS ratings, because they prove a data center’s capability in daily operations.
Within each tier, the Institute is ultimately interested in gauging a facility’s capabilities in terms of:
- Crisis management if some components were to fail
A data center with a robust cooling setup, plenty of power redundancies, streamlined maintenance that doesn’t require much downtime, and a detailed disaster management protocol will then receive a better rating than a facility that struggles in one or more of those areas.
With those key aspects in mind, let’s take a look at the data center tier classifications in more detail.
The first tier covers data centers that can meet basic needs within the four categories listed above. A Tier I center may have a few redundancies capable of safeguarding against mistakes, but they might have trouble dealing with major issues such as outages, equipment failure, and maintenance that requires extensive downtime. Some Tier I facilities offer no redundancies at all.
The typical Tier I data center setup includes:
- Power: A UPS that can handle small dips and surges in power or a slightly larger interruption in power, with a backup generator for longer outages.
- Maintenance: An IT department with qualified staff that can fix problems and plan for preventative repairs.
- Cooling: A cooling solution that remains uninterrupted and operational all day, every day. At a Tier I level data center, this is usually an air conditioning unit.
- Failure Management: At a Tier I level facility, the IT staff may be sidelined by unexpected failures or unplanned maintenance that requires equipment shutdowns.
While not robust enough to service large clients or multiple clients at once, Tier I facilities are sufficient for small to medium-sized businesses that need more power, storage, or network capacity than their in-house IT department is able to provide. Consider a Tier I facility one step up from an on-site server room.
The benefit of a Tier I data center is usually the price point. Smaller businesses may have a tough time affording a data center in a higher tier. With an average annual downtime of 28.8 hours or fewer, however, Tier I data centers may not be the ideal choice for companies that need high availability.
A data center with a Tier II level has built some safety and redundancy into their systems to minimize the risk of faults and downtime. This is reflected in their power and cooling systems, which have a direct positive effect on a facility’s maintenance and failure management:
- Power: Extra generators, including spare fuel tanks with enough capacity to keep generators running longer. A facility may also have the ability to switch freely between engine generators and the power grid.
- Cooling: Upgraded cooling units beyond simple air conditioners.
Tier II data centers are a good choice for small to medium-sized businesses wanting extra assurance that downtime and unforeseen failures will be less of an issue than with a Tier I facility. The peace of mind costs a little more, but if a business may lose significant income during periods of downtime and maintenance, the trade-off may be worth it. However, the annual downtime of 22.7 hours or fewer is still not optimal, and the risk of a total power failure is still relatively high.
Higher levels of redundancy mean a stronger ability to withstand power failures, equipment faults, and human mistakes. By the time a data center reaches Tier III, it should have enough redundant systems in place to nearly eliminate the need for total shutdowns to deal with problems. Maintenance, too, can occur without much downtime, because each major system should have one backup in place ready to take over if need be. While upgraded cooling systems also matter, it’s the backup power and equipment components that separate a Tier III facility from Tier II and below.
Clients who contract with a Tier III data center can expect to have less downtime, if any at all, than they would have in lower-tiered facilities. The official annual limit for acceptable downtime is 1.6 hours or fewer. Additionally, power backups are abundant enough to keep everything up and running for 72 hours when there is no access to the electric grid.
This tier is a great option for medium to large-sized companies whose businesses rely on sustained availability.
Continuing the pattern of advancement, Tier IV data centers are able to withstand a variety of potential catastrophes through the use of multiple redundant systems. At a limit of 0.4 hours per year of downtime, Tier IV facilities should be available 99.999% of the time.
To accomplish this, data centers at Tier IV have backups in several areas:
- Electrical systems
Each redundancy is as isolated and independent as possible to provide standalone support when other systems have failed. Speaking of failure, these facilities are able to withstand power outages of up to 96 hours.
High-security, high-stakes companies need a Tier IV data center. If an interruption of service—including planned downtime or an emergency failure—would spell disaster for the business, opting for a lower tier with less redundancy is an unwise choice.
However, even highly redundant systems are not immune to failure. If one system goes down while another is offline for repairs, there could still be interruptions in service. If this small level of risk is unacceptable for a business, they have other options.
Extra TCOS Levels
Some facilities choose to go above and beyond the requirements that the Uptime Institute sets down for each tier certification. In these special cases, you might see a facility certified as bronze, silver, or gold, signifying their efforts to provide the most bulletproof platform availability possible.
While these tiers would be overkill—and probably far too expensive—for a business in need of a simple data center, they are helpful for hospitals, national security organizations, international financial institutions, and other companies that need 100% availability with a miniscule risk of failure.
Interested in getting your data center certified?
It’s important to remember that the data center tiers explained above are not set in stone. If you own or operate a data center and make significant upgrades, you can achieve higher tier levels. Don’t be intimidated by the thought of bringing in new, upgraded equipment just because the components can be heavy, delicate, and difficult to move. You’re not stuck! Simply rent or purchase a ServerLift data center lift and you’ll be able to rearrange and upgrade your facility when you need to.