Evolution at the Edge

At Dell Technologies World this year, customers and journalists were curious about trends I am seeing in the marketplace and predictions for the future. I shared my views on the impact of 5G , how AI and IoT are continuing to intersect, and the need for businesses to have consistent, flexible infrastructure to quickly adapt. I also emphasized that the foundation of all these transformations is the shift to edge computing—and it’s our OEM & IoT customers across all industries who are leading this evolution.

Location, location, location

At this point, I should clarify what I mean by the edge. I’m talking about data being processed close to where it’s created, versus the traditional centrally-located data center. I like to think of the difference between the data center and the edge as the difference between living out in the suburbs and living in the city—where all the action is. Right now, about 10 percent of enterprise-generated data is created and processed outside a traditional centralized data center or cloud. However, by 2023, Gartner predicts this figure will reach 75 percent. That’s a dramatic shift by any definition.

Three whys

So, why is this happening? Three reasons. First, according to the latest research, the number of connected devices is expected to reach 125 billion by 2030, which will put about 15 connected devices into the hands of every consumer. It simply doesn’t make sense to move all that data to a traditional data center—or even to the cloud.

The second reason is cost. It’s naturally more cost-effective to process at least some of the data at the edge. And third, it’s all about speed. Many use cases just cannot accept the latency involved in sending data over a network, processing it and returning a response. Autonomous vehicles and video surveillance are great examples, where even a few seconds delay could mean the difference between an expected outcome and a catastrophic event.

Edge computing examples

And what kind of compute exists at the edge? Well, it helps me to visualize the edge as a spectrum. On the right end–what I call the far edge–is where data is generated. Picture millions of connected devices generating a constant stream of data for performance monitoring or end user access. One example is a fluid management system, where valves need to be automatically opened or closed, based on threshold triggers being monitored. If this is something that interests you (using IoT data to help customers better manage and troubleshoot control valves), I recommend looking into our joint solution with Emerson.

Or, consider how the frequency of fridge doors opening in the chilled food section of a supermarket affects the fridge’s temperature levels, and ultimately the food. It would be crazy to send to the cloud such a massive amount of data simply indicating the binary safe/unsafe temperature status—the store manager only needs to know when the temperature is unsafe. So, the edge is the obvious choice to aggregate and analyze this kind of data. In fact, we’ve worked with a major supermarket retailer to implement refrigeration monitoring and predictive maintenance at their edge. Today, their cooling units are serviced at the appropriate time, and they’re saving millions of dollars in rotten food. If you’re interested in using data to help avoid food waste, check out our joint solution with IMS Evolve.

Application-driven solutions

Of course, in the vast majority of cases, the application determines the solution. For example, speed in surveillance systems is critical, when you’re trying to find a lost child in a mall or identify and stop someone that is a known security threat from entering a football stadium. The last thing you want at the crucial moment is for a cloud environment to tell you that it’s busy searching.

Thanks to the advent of 5G, carriers are addressing the need for higher data traffic performance by placing servers at the base of cell towers instead of at a regional data center. These are all examples where configuration capability, great graphics and high processing performance come into play. And this brings me to another interesting point. When edge computing started, dedicated gateways were the focus. While still important, that definition has expanded to include servers, workstations, ruggedized laptops and embedded PCs.

The micro data center

Another category of edge compute is what Gartner calls the Micro-Data Center. Many of the attributes of a traditional data center come into play here, such as the need for high reliability, ability to scale the compute as needed, and high levels of management. Conditions that don’t typically demand ruggedized products, but where space limitations are likely.

In these scenarios, customers typically consider virtualized solutions. Remote oil rigs, warehouse distribution centers and shipping hubs are great examples. Just think about the speed of packages flying down a conveyer belt at a distribution center, being routed to the right loading area while the data is being logged in real time for tracking. Batch files are then sent back to a central data center for global tracking, billing, and record keeping. In effect, you have a network of micro data centers at the edge, aggregating and analyzing data, while feeding the most relevant information into a larger regional center.

Looking ahead

In addition to all the practical benefits (such as faster speed, and lower cost), the edge is also driving fresh innovation. After all, the ability to glean immediate insights, experiment, respond in real time and deliver services on-demand are all important criteria in our ever-changing world. In my view, this dynamic will only accelerate with the advent of 5G. By increasing the speed of data analysis, 5G will inevitably increase edge adoption while, on the other hand, businesses using edge computing will experience the full benefit of 5G networks. Over time, I believe that this combination will inspire a slew of new and exciting applications for both the business and consumer markets.