ThorstenSchmitt - Fotolia

News Stay informed about the latest enterprise technology news and product updates.

Tier IV data center comes online as new certification players emerge

New options for data center certification, including UL's debut and the Tier 5 standard from Switch, aim to more easily prove data center resiliency and reliability.

Two new tier certification programs that grade data center facilities will challenge longtime incumbent Uptime Institute LLC at a time when customers demand highly reliable data center capacity, along with a succinct explanation of the infrastructure behind it.

The Uptime data center tier standards are a methodology widely used to explain the resiliency of data center infrastructure. They are often included in a company's request for proposal prior to renting colocation space or building a new data center. Standards range from a Tier I data center with nonredundant systems to a fully fault-tolerant Tier IV data center, the second of which in the United States will open later this summer.

The issuance of tier standards has tripled in the past five years amid surging data center capacity demand -- Uptime issued its 1,000th tier certification last week, and another 419 projects are in queue, said Lee Kirby, president of Seattle-based Uptime. Enterprises that expand their footprint use tier ratings to ensure consistency and reliability across their global data centers.

"More and more companies are counting on the reliability of their infrastructure, and they need to be able to treat it as a utility," Kirby said.

Tier data center competitors

Now entering the data center certification arena is UL LLC. Previously known as Underwriters Laboratories, UL has a long history as a nonprofit testing and certification organization, and it became a for-profit corporation in 2012. Its certification mark is on 22 billion products in the U.S., which range from electrical equipment to smoke detectors, as well as a modular data center safety certification. UL's new international data center certification program, UL 3223, will address operational risks around data center infrastructure, such as fire safety, life safety and security.

A third company, called Switch, claims its new Tier 5 standard evaluates data center resiliency and redundancy, plus other factors such as long-term power capabilities, available carriers, physical security and network security, and use of renewable energy. Switch runs colocation data centers in Georgia, Michigan and Nevada.

The two new data center certification standards reflect the fact that the data center industry wants an easier way to categorize facilities, according to one expert.

There is still clearly a need to describe in simple terms, without looking at drawings, how a data center provides maximum uptime and availability.
Zahl Limbuwalafounder, Romonet Ltd.

"The challenge the industry has -- whoever's tier standard you look at -- is that there is still clearly a need to describe in simple terms, without looking at drawings, how a data center provides maximum uptime and availability," said Zahl Limbuwala, founder and executive director of Romonet Ltd., which provides data center analytics software.

There have been several attempts in the past 20 years to replace Uptime's four tiers, such as Tier III plus or Switch's Tier 5, but these can add confusion over tier standards -- often only for an organization's own benefit. Limbuwala said facility availability and redundancy at Switch are "probably the best [he has] ever seen." But the company's creation of a "Tier 5" certification also lets it claim it has the first tier-five data center, he said.

Pros and cons for top-tier data centers

The majority of Uptime tier-certified data centers are Tier III, but new Tier I and Tier II data centers are now designed for cloud computing and hyperscale data center operators, Kirby said. Most enterprises can't tolerate a data center that is anything less than 100% available because the applications are not set up to failover automatically. Banks and telecommunications companies are most likely to want a Tier IV data center, and some countries, including Italy and Indonesia, have several Tier IV data centers.

Ofek Technologies, a boutique cloud computing company with two data centers in Israel, is in search of data center space in the United States and will only consider Tier III or higher data centers to meet its uptime requirements, said Nir Ashkenazi, the company's CTO.

Data Shelter LLC, in Fort Pierce, Fla., has just completed a Tier IV data center that uses Uptime's certification. The facility sits in a tornado- and hurricane-resistant former nuclear bomb shelter that is mostly underground, with 2-foot-thick rebar walls and a 9-foot-thick floor. A portion of Switch's SUPERNAP Las Vegas campus is also certified as Tier IV.

Data Shelter will target healthcare and financial services companies, government agencies and government contractors, and cloud and managed service providers.

For years, Uptime has battled data center companies that obtained Tier IV certification for a data center design, but later built it to less than Tier IV standards -- but still marketed it as a Tier IV site.

"We always had the vision of a Tier IV data center; we thought it was worth it to bring it to that level," said Data Shelter CTO Mark Oxley. "We now know why a lot of people don't do it -- it was quite an undertaking."

Robert Gates covers data centers, data center strategies, server technologies, converged and hyper-converged infrastructure and open source operating systems for SearchDataCenter. Follow him on Twitter @RBGatesTT or email him at

Next Steps

Discover the latest research from Uptime Institute

Explore common causes of data center outages

How to make data centers more reliable for IoT

Dig Deeper on Data center budget and culture

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

How do you use the Uptime tier certification process when selecting a data center?
I am amazed at the technology and effort to create high availability data centers. However, all too often failures orginate associated with the services leading to the facility. In example, in the article a call out is made to a bunker in Florida. Great, but how will staff and services get to that facility in the event of a hurricane. Roads will be clogged or shutdown, have rain and flooding. I had a backup facility in the middle of Florida that was an ex-army bunker. It was impregnable. However, during a hurricane when we tried to get there main roads were blocked with evacuation traffic and staff did not want to leave families. You have to evaluate these factors as well and all too often they are forgotten or ignored because of the belief you can manage the operations remotely. Really, let me see you get you fuel tanks refilled remotely, Network facilities 20 miles away serviced remotely (you think dual feeds are real then you have never visited a TELCO servicing facility). You data center may be up, but your services may not be able to use them.
Hello Ronolias,

I am the CTO & Founder of Data Shelter in Florida. I fully understand your concerns concerning resupply and access after an event. As a company we are pushing the limits far beyond just the Tier IV rating to look at all of the systems involved with operating a data center with and "island mentality". For instance: We have spec'd an on-site fuel supply of 7 days at full load, our own fuel tanker, a snorkle equipped hmmwv (Humvee), and as luck would have it, have three of the largest refueling stations within 1 mile of the facility. Furthermore the facility sits (west side of the property) on a major highway (FL Turnpike), and is 1 mile from I-95. Our facility also sits in one of the highest land points in the region. We have also invested heavily into the life support systems at the facility. We have emergency CBRN air filtration systems in a positively pressurized building, and in addition to city utilities, have dug in our own deep well with water filtration and septic system and a tertiary 500/gal water reserve tank. Our operational plans include onsite food and supplies for our maximum building occupancy for 10 days. As for Telco, I agree, there is only so much one can do when subject to an internet connection from a local CO that may be affected, but by having multiple fiber carriers, and alternate connections using low orbital satellite and microwave systems, most of the regional issues can be mitigated. There are many more things we have addressed like water-free redundant fire systems in critical spaces, redundant security and access control systems, etc. The goal of our project is currently a proof of concept, but our plans will be to eventually control the networks and power systems that feed them, eliminating the very issues that may stem from the questions you posed. I appreciate the experience you have shared and hope you can see that we are trying to look forward beyond network, electric, and cooling redundancy, and into all of the ecosystems that make up a data center.