Finding Strength in Numbers: The Data Center Clustering Effect

Loudon Blair is Senior Director of Corporate Strategy for Ciena.

Did you know that the best place to put a new Burger King is next door to a McDonalds? Have you ever noticed that certain types of retail stores tend to cluster, and have locations near each other? Significant time and money goes into picking the most strategic location for a business, so there is little chance this is a simple coincidence. The same can be said of data centers. More and more of them are popping up near one another, forming clusters around the U.S. - but why?

Historically, data centers had been located in remote locations. Cities were simply too expensive – from the land to energy costs. When the biggest priority was lowering the price per kilowatt, cities simply could not compete. This all changed, however, when the competition started heating up.

Rising customer expectations and increased bandwidth requirements have encouraged data centers to move closer to the end user, and with an increase in the number of applications and services that require real-time delivery of data, this trend will only continue. From virtual reality to autonomous vehicles and streaming services, even the smallest amount of lag in network performance can have a big impact on experience; and, therefore, the bottom line. If you look at the world of augmented reality (AR) and virtual reality (VR), for example, “the goal of this medium is ‘true immersion’ — creating an illusion so real that it convinces the human brain, the world’s finest computer.” This is according to a whitepaper titled, “Graphics Processing Requirements for Enabling Immersive VR” by D. Kanter, which was featured in a report from AT&T Foundry. The report goes on to say that the “general consensus is that target response latency for both AR and VR systems is easily three to 10 times lower than for today’s standard non-HMD [non-head-mounted display] visualization.”

The push to build data centers closer to metro areas such as: New York City, Los Angeles and San Francisco where there is a large concentration of customers represents one of the first examples of the data center clustering effect. Specifically, data centers clustered at existing peering points through which all customers’ traffic flowed. Then, as cloud computing gained momentum, data centers moved to locations to where their businesses were located. The fact is, though, a large portion of providers’ customers and businesses are located outside of these large cities. That has driven the move toward building data centers in secondary, and even tertiary markets.

Plus, with advances in data center interconnect (DCI) technologies, data centers don’t need to be in the same city to achieve a smooth transfer of critical assets. The technology makes it possible for physically separate data centers to easily share resources and balance workloads.

Getting Closer to Local Communities

One of the most recent high-profile acquisitions - Amazon’s takeover of Whole Foods – may not appear to have anything in common with data centers, but in fact, there are some stark similarities. One of the reasons Amazon purchased Whole Foods was for its network of local stores. Amazon wanted to get closer to its customers so they could reduce the time it takes to deliver goods. Similarly, in communications, short delays (or latency) leads to happier customers.

The other part of the equation is the type of data that is processed in these local data centers. There is certain data that is only relevant to the local community, like navigation, certain voice-recognized questions or targeted ads, which all need to be provided with incredibly low latency.

Local Distribution, Local Resources

Now we know why data centers are moving, but why are they moving together?

Foundation is created: When one data center moves in they do a lot of work up front - everything from negotiating tax incentives to building much needed resources like cooling systems. Many times, after this foundation has been created, other data centers will come to town as most of the heavy lifting has been done for them. For example, there are tax advantages when real estate investment trusts (REITs) are established, and lower costs when power and communications infrastructure are already in place. Once the foundation has been established, data center operators swarm to an area like bees to honey.

Market on board with multi-tenant data centers: Ten to 15 years ago, people stayed away from multi-tenant data centers. They didn’t want someone else in charge of their data center – simply too much was at stake. Today, people take it for granted, especially as more applications move into shared ecosystems.

Lack of space in major city hubs: Space is limited, and data centers can get pretty big. In fact, the world’s largest center is being planned now – it will “eventually draw on a record-setting 1000 megawatts of power,” according to a recent CNBC article. When companies look to build a new data center, and space is a concern, they will look to other areas that are in close proximity to these larger cities. These space constraints affect everyone, and the number of locations around these metro cities are finite.

Finding Strength in Numbers 

As we’ve seen with fast food restaurants and gas stations alike, there is a benefit to sticking together. The same is true for data centers. With data centers storing and protecting information –  the lifeblood of most companies – operators are going to continue to look for any advantage they can find.

Opinions expressed in the article above do not necessarily reflect the opinions of Data Center Knowledge and Informa.

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.