As you begin to learn more about data centers, you might notice that they all claim to have some sort of reliability certification. This data center tier certification is managed by the Uptime Institute. It is the industry standard for data center uptime certifications around the world.
The Uptime Institute designed the tier system to be broad in order to accommodate a wide variety of data centers: design, locations, local codes, etc. Rather than holding data centers to rigid technical requirements, the tier system allows data center owner/operators the freedom to innovate, while still assuring their customers a level of reliability and uptime.
The chart and explanations below define the Uptime Institute’s requirements for each tier.
Number of delivery Paths
Delivery paths are the primary electrical circuits that power the data center. Tier I and Tier II data centers only need to have a single electrical path into the facility. Tier III and Tier IV data centers need to have redundant electrical paths to ensure that if one path fails, the second can continue to provide electrical power without incurring any downtime.
Redundancy can relate to all the core components of a data center. An example using backup generators best illustrates this metric:
Let’s assume that 5 backup generators are needed to ensure uptime in case grid power gets compromised. The 5 generators would meet the N rating for a Tier I data center. If that same data center had 6 backup generators, it would meet the N+1 rating for Tier II & Tier III facilities. In order to meet the 2N Tier IV standards, that data center would need a minimum of 10 generators. Some data centers even advertise that they have a redundancy of 2N+1, which would mean double the necessary backup generators plus at least one more.
The same redundancy calculation can be performed for any of the mission critical components of a data center: power systems, cooling systems, backup generators…etc.
Only Tier IV data centers are required to be compartmentalized, meaning that each of the power systems and paths from one piece of equipment to another have been isolated from each other. This ensures that each individual part of the infrastructure can function independently from the other machines. So, if one machine goes down, it does not affect the other machines.
Tier III and Tier IV data centers are required to be concurrently maintainable, meaning that maintenance can be done without disrupting IT operations. This allows for each piece of equipment to be removed, replaced, and serviced without incurring any downtime.
Only Tier IV data centers are required to be completely fault tolerant. This means that the site can withstand natural disasters, blackouts, and other catastrophic events that would cause downtime for a normal data center.
Availability / Downtime
Availability and Downtime is the same metric expressed two different ways. Availability is the % of uptime maintained in a year. Downtime is the time, expressed in hours, a data center is allowed to be offline (here is a nice summary from Colocation America).
For example, a Tier I data center is allowed to be down only 0.329% of the year (or 28.8 hours per year). You might be wondering that even the lowest-rated data centers have exceptionally high standards. This is because even a few hours and even minutes of downtime can add up to a lot of lost money. The most recent survey from Emerson Power Network and the Ponemon Institute found that the average unplanned downtime cost for businesses was $8,851 per minute!
How do people decide which Tier to conform to?
Tier I and II data centers are usually meant for smaller businesses that are driven by cost over reliability. Whereas Tier III and IV data centers are meant for larger enterprises that are more concerned about continuous availability.
However, that does not mean that a Tier IV data center is a better choice than a Tier I rated data center. In the end, it all depends on your needs. A Tier IV data center might not be as cost effective if you are running non-critical workloads. On the other hand, if you are require a continuous availability, a Tier I facility might not be reliable enough for your needs.
The Uptime institute gives out two certifications in order to ensure that data centers continue to meet their standards. They are the Tier Certification of Design Documents (TCDD) and the Tier Certification of Constructed Facility (TCCF).
Data centers can receive a TCDD certification based on the design blueprints for the data center. However, that rating may be adjusted in the construction phase and after the facility is fully built. So, it is important to make sure that a data center meets the required criteria after design and construction.
The Uptime Institute is trying to clarify some confusion surrounding these two milestones. They have said, “Uptime Institute is addressing the concern that a Tier Certification of Design Documents could be used to substantiate a data center that is designed to one Tier level and constructed and commissioned to another Tier level.”
A data center can only obtain a TCCF certification when consultants from the Uptime Institute physically visit the data center after is has been built. While they are there, they can determine if the data center’s blueprints actually meet the design specifications.
If you want to check on a data center’s certifications, you can find a map of all data centers that have been certified by the Uptime Institute here.
This article was researched and written by Clyde Stanhope.