With degrees in marketing, economics, and psychology, I have taken more survey research and statistical analyses classes than anyone should ever have to do. Still, truth be told, I love a good survey. The nerd in me gets a little bit giddy knowing that my anonymized data is going to be used to drive some change in the world because I did my part to weigh in when I was asked. But as any researcher knows, the more people you ask a question to, the greater distribution of responses you will receive.

If you asked 100 people what an edge data center was, you could easily have 100 different answers. Most would boil down to, “It’s a small data center located closer to end users that reduces latency.” But ask if they can point to one, and they might side-step and quickly excuse themselves. 

I have worked for companies rolling out an edge cabinet solution (that didn’t really perform well) and I have worked for others conducting hefty research and analyses of the edge market so they could have a first-mover advantage with a new offering (that they ultimately decided was too risky to launch). These were really stable companies that were otherwise hugely successful in the industry. I share this because, when it comes to the edge, many have taken a “build it, and they will come” approach and failed.

What we haven’t done is reorient ourselves to the ways the edge is already being rolled out and try to map our plans to it.

We operate in a world of standards, of reliable and predictable uptime, of engineered perfection. The indefinite has no place in our dictionary. Except that it must. I would suggest that, perhaps, we aren’t asking the wrong question, but we’re asking the wrong people — namely, we keep asking ourselves and the answers keep coming up short. Over the years, I’ve heard statements and questions like the following.

• “I need to definitively say whether this sale is to a colo or an edge data center.”
• “That isn’t the edge; it’s modular.” 
• “We already support telecom; the edge play can’t be the same.”

This survey enthusiast wants to slam her fist down and yell, “It’s D, none of the above!” Truthfully, none of these are wrong, because the edge has less to do with a physical location or description, and is based more on what it is doing.

To take you down memory lane, about a decade ago, I was between positions, long before I would consider myself having been in the data center industry. At the time, I had young kids and an intense disdain for red-light runners. It frequently crossed my mind that I should change my profession to become a traffic cop, with the reality of the situation being that no one could possibly take someone of my size or demeanor that seriously. But then a friend from a former software startup called and asked if I’d like to take on some side work at his “server storage facility” reviewing video footage for one of the clients leasing space. It was red light traffic camera feeds that needed human verification for accuracy. In a digital way, I could make my dream come true! 

Why does this matter, and what does it have to do with the edge? Well, these were localized feeds, triggered by smart sensors throughout the city being fed into the system in real time. This wasn’t the entire department of transportation or even my entire state. The data was specific to the Central Florida area, where it wasn’t just going to be screened by me for verification, but assumedly by the individuals who would be issued a ticket for their offense so that they could review the footage and either pay their fine or contest the incident, all within their local jurisdiction. This was ahead of the times! Not only did that video footage not need to be transmitted across great distances for me to add my stamp of approval, but the driver’s financial or legal response didn’t need to route to a centralized DoT server either.

I use this as a prime edge example because it was built and operated entirely around a use case to process and manipulate data as close to both the originating source and end user as possible. It had nothing to do with how the data center wanted to operate. It was a revolutionary example of a client, the DoT, knowing their users (drivers) well enough to plan around their needs. With the focus on reduced latency, backed by intimate knowledge and proximity to the end user, we could change how we delivered solutions based on these needs.

The edge can be a colocation, it can be modular, it can be in a telecom tower, it can be a cabinet in a manufacturing plant. It can be all these things, because it has absolutely nothing to do with us and how we want to build it. Moreover, we cannot predict what applications are going to hit the market with their own unique demands. From gaming and autonomous vehicles to generative AI and the work-from-home shift, people are immersing themselves in new digital experiences that we need to quickly understand, so we can identify the best ways to collect and distribute content as close to each source as possible — as quickly as possible. 

Each use case will define what the ideal looks like for that scenario. In this case, everyone having their own definition of the edge is actually a good thing. The edge is everywhere, and I think that’s where it’s supposed to be.