Unless you’ve been living under a rock for the past few years, you know that you can’t swing a dead cat nowadays without hitting someone talking about the “edge.” One could make a serious argument the edge has supplanted the cloud as the IT topic that everyone talks about but doesn’t agree on a shared definition. Of course, these definitional arguments can go on for years, but the bigger question in search of an answer is “What’s driving everyone to the edge?”
From an enduser perspective, the demands for faster data processing and delivery of content and services to mobile devices will be unceasing. The projections are literally at a steeper angle than the growth of our national debt. According to Ericsson, there will be 6.1 billion smartphone users by 20201 a figure not wholly unexpected based on their current degree of ubiquity, the fact that 10.2 million units of “smart clothing” will ship in that same period, however, probably isn’t on anybody’s radar2. We’ll also be able to track pretty much anything in the coming years as current estimates say that there will be over 55 billion Internet of Things (IoT) devices and over 250 million vehicles will be connected to the internet3 by 2025 per Gartner. All of that would be well and good if the current structure of the internet and cloud and mobile networks were able to keep pace. But they can’t. Not even close.
But don’t take my word for it. Just look at the latency, because it speaks for itself. And it’s crying “uncle” under the pressure already, even though the demands of edge applications are just getting started. The centralized data center model that dominates the internet’s current infrastructure means that content and services live a vast distance from most users, having to travel a gauntlet of delay-filled obstacles as bits and bytes travel to and from mobile devices. Simply put, the latency levels associated with the current structure of the cloud are impractical to support the immediacy requirements of many new applications that will increase in the next three to five years, including:
Virtual (VR) and augmented reality (AR)
Connected and driverless cars
IoT sensor networks
Smart Cities applications
Next-generation gaming and mobile entertainment consumption
The intended purpose for virtual and augmented reality is the delivery of the most realistic end user experience possible. Coupled with devices such as Oculus Rift or Samsung Gear VR, mobile phones deliver an immersive experience to the enduser. The level of detail and functionality conveyed by these offerings and those that emerge to compete with them will only continue to become more elaborate and, correspondingly, require higher bandwidth and faster delivery to accomplish their mission.
While gaming has been the predominant use of VR and AR to this point, the use cases within industries such as health care, construction, and education will exacerbate the need to move “cloud-like” functionality ever closer to the users of an application be it a hospital or a construction site. These industrial/commercial applications of VR and AR are projected to see enormous growth, and they simply will not be possible without edge infrastructure that lowers latency significantly.
Connected and driverless cars
Although developing cars that can safely operate without a driver is taking a little longer than many proponents have predicted, when they do arrive en masse they’ll be generating a lot of data. Hitachi Data Systems estimates that a single vehicle will create more than 25GB/hour.
While these vehicles will have onboard computers, edge connectivity will be essential in many ways. For example, edge facilities could be used to analyze vehicle-generated data to identify and predict problems to provide for more-timely and less-disruptive preventive maintenance. Other data generated can be used to investigate things like traffic flow by daypart to enable more effective road design and traffic light timing to make everyone’s morning commute just a little more pleasurable.
Internet of things
Although we’re probably a few years away from the state of nirvana when your refrigerator proactively informs you that you’re almost out of beer, there are already many consumer-focused IoT applications in use right now. My new washer and dryer for example, are sending maintenance-related data to the manufacturer. They still lose my left socks on a regular basis, but I doubt there is an IoT solution for that.
The industrial/commercial side of the IoT world is where the growth is already explosive, with billions of sensors and connected devices in the process of being deployed. The demands of these devices alone are making some experts describe IoT alone as the dominant driver for edge computing — making edge infrastructure a must-have for performing tasks such as protocol translation and data transformation as well as housing the AI models that provide the intelligence to connected devices.
In the world of Smart Cities, everything from your home, its meters and controls, grids, buildings, etc., are consistently operating based on the data they generate. The data volumes created eliminate centralized data centers or clouds from being able to do the high-level processing of information that will control everything from traffic lights to building temperatures. As, a result edge data centers will be located throughout the municipality, and serve as the primary point of information processing while forwarding large chunks of data to the cloud for storage and higher level processing.
Gaming and mobile entertainment consumption
Whether it’s playing Fortnite, or whatever eventually renders that passé, or watching Game of Thrones from your dentist’s waiting room, you’re using a lot of bandwidth. In both instances, user satisfaction is predicated on the bi-directional exchange of large volumes of data. Edge computing is essential to delivering the lowest possible levels of latency and efficient bandwidth utilization, a trend that will only continue. The future of entertainment demands edge computing infrastructure in order for consumers to be satisfied with the user experience.
The future needs the EDGE
Currently, edge computing is in its infancy. However, the nature of applications development in both hardware and software points to its inexorable spread throughout the next three to five years. Currently, implementations tend to be small and driven by particular organizational requirements. For providers, business models vary based on vision with some focused on the deployment of large scale networks consisting of collocating their edge computing facilities at the locations of significant cell tower providers.
Similar to traditional retail colocation, space in these sites would then be leased to one or more endusers. The scope and availability of collocated units will be a function of both the availability of capital to fund implementation efforts and the location of the networked locations. The other pole of provider models is based upon providers being enduser agnostic as to their customer base, thereby offering endusers the immediate availability of edge data centers for their network implementations. The one sure thing about the movement to the edge will be that it will continue to move closer and closer to the users it’s designed to serve in the years to come.
3 Business Insider Intelligence