As a designer of mission critical facilities, I’ve learned that it’s really difficult to build data centers to keep pace with technology, yet that’s a reality we face along with our clients, writes Gensler's Jackson Metcalf.
Image: courtesy Gensler
As iPhone 6 mania continues to sweep around the globe this week, and people are newly focused on how mobile technology will transform their lives, I’m thinking about the data centers that underpin our computing systems. When our use of technology grows and changes, the infrastructure that supports it changes, too.
As a designer of mission critical facilities, I’ve learned that it’s really difficult to build data centers to keep pace with technology, yet that’s a reality we face along with our clients. At Gensler, our Mission Critical practice watches changes carefully. Here are the four trends we see impacting data centers most today.
1. More is more. Our need for mission critical facilities is growing, and the way that enterprise companies are approaching the way they build data centers is changing, too. Rather than focusing on single, massive facilities with layers of redundancy and hardening, we’re seeing clients create multiple data centers in different geographic areas. They’re doing this in different ways, to suit their distinct needs.
Some clients create both primary and disaster recovery facilities, while others pursue dual primary facilities with the capacity to back each other up. Another approach is to create a primary facility, and leverage the cloud for back-up, as the Department of Defense announced it’s doing this week. The bottom line here is that there is benefit in multiple facilities, in different geographic locations. This enables clients to have data centers where they need them, and minimize risk by creating facilities in different locations.
2. Tier 1 is the new Tier 4. As our clients create a greater number of mission critical facilities, those that they do create are more efficient, minimal, and nimble. They’re less precious, and as a result, less costly. For several years now, we’ve watched the major tech companies break down their mega data-centers, opting instead to create agile networks of interconnected facilities. This enables tech giants to re-route traffic globally.
If one geographic area goes down or becomes bottlenecked, they are able to re-route traffic to their other facilities. The result is that Tier 4 facilities—those with the highest rating of the Uptime Institute—are falling out of favor as companies look to locate multiple, smaller data centers—Tier 1 or Tier 2 facilities—in different locations. While this trend is well underway with tech giants, we think it will soon trickle down to enterprise clients and typical business users who operate their own facilities.
3. Shorter lifespan for facilities. It’s exceedingly difficult to build mission critical facilities for technology that has not yet been invented, yet that’s the challenge our clients face every day. Because tech life cycles can be as rapid as 18 months, and it takes that long to build a data center or even longer to plan one, we’re constantly challenged to design for the unknown. Additionally, as ever-more-efficient mechanical and electrical equipment comes to market, we see many opportunities to improve data center efficiency on an ongoing basis.
A typical building’s electrical system might have a 20 year usable life, but for data centers, those systems could be replaced long before the end of the system’s useful life in order to reduce operational expenses. As a result, clients are opting to minimize investment in “Swiss Watch” style facilities, opting to plan simpler, more flexible facilities instead. Ten years ago, we built many mission critical facilities using 50-100 year construction methods. That just doesn’t make sense today. As rates of change for both technology and equipment accelerate, our clients want to be able to easily change out a building’s mechanical and electrical systems—not just their server architecture.
The challenge lies in creating facilities that remain relevant, and our tech clients are doing this particularly well. They’re building data centers in a prototypical, modular fashion that enables them to grow and change out mechanical and electrical systems efficiently.
4. Plug & play. For data centers, rapid change is inevitable as facilities scramble to keep pace with tech cycles. As a result, our clients have always designed ways to create modular, scalable computing systems. Today, we’re finding ways to do the same thing with the buildings that house their computing systems.
For example, there are two primary ways to cool a mission critical facility: with water cooling, or by using outside air. I’ve seen a client create a day-one system designed to use outside air, yet we embed flexibility in the design so that if in five years that system doesn’t make sense, the client can quickly remove a wall of louvers, put up a solid wall, and install cooling towers. We design facilities with that kind of modularity so that as new systems emerge, our clients can quickly and efficiently decommission and change them.
Another way we’re making change more easy and quick for our clients is by designing buildings using off-the-shelf systems, rather than custom. For data center clients, the ability to shave three to five months off a construction timeline because we’ve used an off-the-shelf switchgear rather than waiting to construct a custom-built one, for example, can make a huge difference. Speeding delivery, and designing buildings so they can also be quickly changed, is key.
About the Author
Jackson Metcalf finds ways to create award-winning data centers noted for their sustainable design and technical prowess. Four of his data centers have achieved LEED Gold certification, which is one reason he’s seen as a leader in our Mission Critical practice area. Contact him at firstname.lastname@example.org.