Finding strength in numbers: Data center agglomeration effect

In the past, data centers were often built in remote areas far from cities because it was expensive to build and operate data centers in urban areas, from land resources to energy costs. When the main priority for data center operators was to reduce electricity costs, it was difficult for cities to have a competitive advantage. However, when market competition began to heat up, all this changed.

Did you know that KFC is usually built near a McDonald's? Ever notice that certain types of retail stores tend to cluster and be located next to each other? Businesses often spend time and money choosing a strategic location, so this is not a simple coincidence. The same is true for data center construction, which is increasingly being built next to each other and clustered across the United States. Why is this?

Rising customer expectations and increased bandwidth requirements are driving data centers closer to end users, and this trend will continue as the number of applications and services that deliver data in real time increases. From virtual reality to autonomous vehicles to streaming services, even the smallest delays in network performance can have a big impact on the experience. For example, if you look at the world of augmented reality (AR) and virtual reality (VR), the goal is "true immersion," which is to create a real illusion that the human brain is the world's best computer. This is based on the white paper "Graphics Processing Requirements to Enable Immersive Virtual Reality" written by D. Kanter mentioned in the AT&T Foundry report. The report states, "There is general consensus that the target response latency for augmented reality (AR) and virtual reality (VR) systems should be 3 to 10 times lower than the target response latency of today's standard non-HMD (non-head-mounted display) visualization."

[[207695]]

Today, data centers are being built closer to urban areas, such as in New York City, Los Angeles, and San Francisco, where there is a large customer base, which is an example of the data center clustering effect. Specifically, data centers are concentrated at existing network peering points where all customer traffic is located. Then, as cloud computing developed, data centers moved to where their business was located. However, the fact that most of the providers' customers and businesses are outside of these large cities has driven data center operators to build data centers in secondary and even tertiary markets.

In addition, with the development of data center interconnect (DCI) technology, data centers no longer need to be in the same city to achieve the transfer of key assets. This technology enables physically separated data centers to easily share resources and balance workloads.

Close to local communities

Amazon's acquisition of the Whole Foods supermarket chain is one of its recent high-profile acquisitions, which may not seem to have anything in common with data centers, but there are actually some clear similarities. One of the reasons Amazon acquired Whole Foods was because of its network of local stores. Amazon wanted to be closer to customers, so they needed to reduce the time it took to deliver goods. Similarly, in the communications industry, short delays (or no delays) make customers happier.

Another reason is because the type of data processed in these local data centers is relevant to the local community, such as navigation, certain voice recognition problems, or targeted advertising, all of which require very low latency.

Local distribution, local resources

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

(1) Creating the foundation: When a data center moves, the operator and local government do a lot of work, from negotiating tax incentives to building the necessary resources such as cooling systems. Many times, after the foundation is established, other data center operators will come to the city to build because most of the work has been done locally. For example, real estate investment trusts (REITs) companies offer tax incentives when building data centers, and the power and communication infrastructure is already built, which reduces costs. Once the infrastructure is completed and the conditions are met, data center operators will flock to the area to build and operate data centers.

(2) Multi-tenant data center market: 10 to 15 years ago, people stayed away from multi-tenant data centers. They didn’t want someone to control and manage their data center. Today, people take it for granted, especially as more applications move into shared ecosystems.

(3) Lack of space in major urban hubs: Due to limited space in major cities, data center construction can become quite large in order to make full use of the space. According to a recent CNBC article, the world's largest data center is being planned, with a power capacity of eventually reaching a record 1,000 megawatts. When companies seek to build a new data center and space becomes a concern, they will look for areas adjacent to these larger cities. These space constraints affect everyone, and space in these metropolitan cities is limited.

Finding strength in numbers

Just as we’ve seen with fast food restaurants and gas stations, where clustering can be extremely beneficial, the same is true with data centers. As data centers become more important for storing and protecting information, data center operators will continue to look for any advantage they can find.