Assessing the Importance of Data Center Tiers
With more than 16 years of technical and consulting experience in data center migration, design, management and operations, and sustainable practices, Kris Domich is the principal consultant of enterprise services and data center solutions for Dimension Data.KRIS DOMICH
In my professional life, I’m focused on helping enterprise clients build high-performance data centers. Outside of that, I also operate a small commercial, organic farm on which I use no pesticides, herbicides or non-naturally occurring fertilizers, even my tractor is primarily fueled by biodiesel. While my farm meets the majority of the criteria to be certified organic, it isn’t. Why? The reason is simple. The cost, time and effort it would require to obtain certification is prohibitive − especially when I’m still providing my customers with the same or better quality products.
This scenario is similar to one I see on a daily basis in the data center industry – Uptime Institute Tier certification. Significant effort and cost is required to achieve it, but does certification increase uptime or service availability? There are many data center providers that have no Tier certification and still offer the same or better service. For example CDGI (Cyber Development Group International) of Mt. Prospect, IL, a data center service provider that designs and deploys mission-critical facilities according to the published standards from TIA, ASHRAE, NFPA and NEBS, houses some of the IT assets of Chicago’s most prestigious companies and assures zero downtime ─ without Tier certification.
Over the past ten years, I have worked on the site selection, retrofitting, build-out and migration of over three million square-feet of mission-critical data center space. The topic of Tiers comes up in my conversations with nearly every data center client. Usually it is a self-scoring discussion where my client tells me what Tier they believe their data center is, divulges aspirations to achieve a certain Tier, or asks me what Tier I believe they need to hold. My initial response is always the same – “Why do you believe you are or need to be at any particular Tier?” No matter the organization, the reply is almost always rooted in the belief that a given Tier guarantees some level of uptime and/or availability that the organization feels they should have.
A Look at the Numbers
Digging deeper into this discussion, many data center providers and enterprises fixate on the uptime percentages commonly associated with a given Tier − Tiers I, II, III and IV should guarantee no more than 28.8 hours, 22 hours, 1.6 hours and 2.4 minutes, respectively, of annual downtime. This suggests that if a facility is not certified, the downtime metrics may be higher. This presents a perceived problem for many of my U.S.-based clients because according to the Uptime Institute’s website, there are only 38 certified data centers in the U.S. Of these, 23 are corporate data centers and the remaining 15 are a combination of technology and service providers. So, for clients looking to move their critical infrastructure into an outsourced data center with Uptime Institute Tier accreditation, there are currently only 15 data centers from which to choose.
Drilling down even further, this means only 1.6 percent of the 887 wholesale and retail colocation-capable facilities listed on DataCenter9.com’s most recent U.S. data center locater map are certified. The remaining 98.4 percent of uncertified data centers still house hundreds of thousands of mission-critical servers, storage arrays, and immeasurable amounts of data and applications. This is not to suggest that data center outages do not occur. Clearly they do, but the data center itself is not the only variable that contributes to uptime. For example, well-architected services can ride out most facility outages as a result of redundancy, geo-diversity and other well-established best practices. Conversely, poorly architected services in a Tier III or Tier IV facility could easily experience reoccurring outages even for simple maintenance tasks.
Given these statistics, it’s reasonable to question the value of Tier certification – especially when you factor in the substantial financial and human capital investment required, and the additional costs of remediation activities if a data center falls short in its attempt to meet the certification. This also assumes that the remediation activities can be performed without posing intolerable risk or imminent downtime to the facility.
Digging into the Rationale of Tiers
Over the years, I’ve worked with numerous enterprises that were planning to seek Tier certification. Their reasons varied ─ from a marketing differentiator, to peace of mind, to the belief that it is a best practice.
The marketing goal is clear enough. After all, to be one of so few certified facilities means providers should be able to charge a premium for colocation services, right? In theory, this may be the case, but the reality is a bit different. To become Tier certified, you have to go “all-in” – meeting a range of requirements around design documents, the facility and operational sustainability. Each of these is a separate certification and can be issued individually. To complicate matters further, the meaning behind the Tier classifications remains clouded. So, for example, if data center A has a Tier III Certification for its Constructed Facility, and data center B has the same certification plus Tier III for Operational Sustainability, which one is more “available”? Which one is more vulnerable to utility outages, fiber cuts, or lack of ability to procure fuel for emergency power during a geographic catastrophe? Unfortunately, the answer isn’t always entirely clear.
Data center providers seeking peace of mind must consider the perspective of their customers, end users and other stakeholders. These audiences don’t see the data center from the standpoint of whether it is concurrently maintainable or fault tolerant. In fact, most users do not have any awareness of the data center. All they know is that when they click a button, open an application or send an email, these things simply work as they should. The underlying infrastructure – the data center facility and the MEP (mechanical/electrical/plumbing) systems are “black boxes” to most customers and users – they are aware that they exist, but they don’t know what they are comprised of or the relevance to uptime. The reality is, even a Tier-certified facility is only as reliable as the most poorly deployed and maintained application or service. As such, for true peace of mind, I would suggest focusing first and foremost on the service architecture as a driver for data center availability. Applications that can span multiple data centers with built-in fault tolerance are inherently more powerful than a Tier certification.
Lastly, Tier certification is considered by many corporations worldwide to be a best practice. However, as outlined above, it’s important to remember that just because a facility is Tier certified, it does not necessarily mean that it is the best of the best.
Buyer Beware: Do Your Research
At the end of the day, just as organic food buyers see value in organic-certified groceries, many organizations see value in the Uptime Institute Tier certifications. Similarly, just as Stanford University scientists revealed in their recent study that organic foods may not actually be healthier for you, having a Tier certification isn’t necessarily guaranteed to help your facility maintain a certain level of uptime each year. As such, when enlisting a data center provider as a business partner, it’s important to simply be alert, and to thoroughly vet the data center provider’s track record and assess the provisions they offer.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.