According to a recent JLL report, edge data centers are predicted to become a $317 billion market globally by 2026. Edge data centers are defined by JLL’s research “as a facility that brings computing power closer to where the data is generated or consumed.” Overall, edge data centers facilitate faster decision-making and more efficient operations. In conversation with Kristen Vosmaer, Managing Director, Data Center Work Dynamics, JLL, he presents an overview of edge data centers, their anticipated growth and how they compare to traditional data centers. Read his below insights to learn more:
The report, Global edge data center market to cross $300B by 2026 found that edge IT infrastructure and edge data centers will become a $317 billion market globally by 2026, representing a 107% growth from the 2020 market valued at $153 billion. Starting out, let's chat about the biggest differentiators between edge data centers and traditional data centers. What makes them different?
We might also use the term “production” data centers when describing traditional data centers, but the way to distinguish the two is by the function. The production data center is the hub for running a query on your search engine or on your social media platform. It’s what people think about when it comes to a stereotypical data center; it provides efficient access to various data sets to pull together the response that you’re querying on. From an edge perspective, however, it’s more purpose built and is typically smaller even down to an IT closet. Edge data centers anticipate what most requests will be for the data center. So, for example, if a new Walter Payton documentary was coming out, the people in Chicago are more likely to watch that documentary, and the content provider may recommend storing the data center locally in Chicago for the viewers to quickly access that data. This avoids frequent use of long haul over fiber which becomes expensive
JLL defines and edge data center as a facility that brings computing power closer to where the data is generated. Another source describes that edge data centers use a smaller footprint and delivers faster services. What other components are important to note about edge data centers?
Edge data centers could operate with just one rack of servers and could be completely self-contained. Meaning, they can be dropped into existing data centers or other types of real estate and that functions as an edge data center. The main goal with the edge data centers is that they direct traffic or quickly recall a certain data set closer to the source. IoT, or Internet of Things, creates all of this data and they may not want to always send it to a remote data center, they’ll want to send it to a closer data center. As a quick aside, Amazon’s simplification for the Internet of Things is the “network of connected devices and the technology that facilitates communication between devices and the cloud.” This can also apply to the billions of everyday items we have connected to the internet.
The JLL North American Data Center report also said that these newer data centers don’t require backup power or as much redundancy. Can you describe that in more detail?
Edge data centers could reduce dependence on backup generation because of their specific r function. In general, backup generation ensures availability of the servers in the event of a grid outage. Newer generations may not require that functionality. For example, an AI workload could hit the “pause button,” like we do when we save a word document mid edit, while batteries temporarily support the load during a grid outage. and then resume once the power turns back on. These backup generation plants are built to support the IT load, as well as the cooling function that’s required to maintain 100% continuity in the data center. By eliminating backup generation, a company saves significant capex and reduce building and operating costs. Also, there’s a massive gain on the sustainability side because it reduces your carbon footprint production.
The report said that in the U.S., 21% of data center development is happening in edge geographies. Do you expect that we will continue to see this steady growth?
I do expect that we’ll continue to see this growth in tier one, tier two and tier three markets and it’s going to be simultaneous. The reason it will be simultaneous is that, when AI infuses more capabilities into applications and technologies, this will then further define what the edge data centers are required to support. Large hyperscale data centers will chase lower power and lower development costs and continue to pop up in these rural areas because in Texas, for example, you can get power sub five cents a kilowatt hour versus California, which is around 21 cents a kilowatt hour. Predominantly, edge data centers will continue to grow around the urbanized areas, but there will also be rural growth for edge compute.
In the report, Global edge data center market to cross $300B by 2026 they said that "The continual growth in data and connected devices has driven the need for increased storage." Can you break down the types of storage needed? For instance, how does the storage needed for edge data centers compare with traditional data centers?
The production or edge data center’s storage will be dependent on the outcome they’re required to produce. One of the driving factors for storage, which impacts edge data centers, is the fiber cost. There is a high cost for that fiber backhaul and pushing data across on a frequent basis. In short, sending data back and forth repeatedly is expensive and takes away bandwidth needed to support traffic to large data centers. At the edge level, there may be a requirement to store incoming data, but choose to store the date until an optimum time to send that data to the production facility, so when there is less traffic on the network. The storage must be sized for these interactions between edge and production data centers.
Another point that was made is that edge data centers are essential for improving security measures and protecting against potential disruptors. How is this possible?
One way to look at this from a cyber security perspective is extending the fence around your property, which creates a larger buffer zone. This buffer limits exposure by processing data closer to the source, which reduces risk. The edge network becomes an extended layer that you’d have to penetrate before you can get to the next layer and have a limited quantities of data compared to large data centers to reduce your exposure.
Interested in more content like this? Read our recent articles, Mid-Year 2024 U.S. Commercial Real Estate Outlook is Mixed Bag or Report Finds Urgent Demand for Jobs Needed to Support Manufacturing Construction. You can view all this content and more when you click here to sign up for our Industrial newsletter.