US telco AT&T plans facility where partners can run edge computing experiments in areas such as self-driving cars and augmented reality.
With edge computing now firmly finding its place in the IoT, US telco AT&T has laid down plans to open an edge computing test zone in the Bay Area of northern California in early 2018.
The zone itself is intended to be a cross between a proof of concept (PoC) lab and a developer hack shop. Initial reports suggest that AT&T will invite partners to test connected applications there, such as self-driving car software, drones and augmented and virtual reality (AR/VR) innovations.
At launch, the zone will use a 4G LTE connection, but the engineering team behind the lab zone hope to upgrade to 5G once the final standards and equipment are ready.
Read more: AT&T fires up LTE-M network in the US
The next step
“Edge computing is the next step in the evolution of the network,” claimed Melissa Arnoldi, president of AT&T Technology and Operations. “As [fast] connectivity becomes ubiquitous, it also needs to become smart. Edge computing puts a supercomputer in your pocket, on your wrist, in your car, and right in front of your eyes.”
The company has suggested that edge computing’s core challenge is striking the right balance between functionality or power. For example, today, an AR app running on a smartphone can offer high-end images or longer battery life, but not both. Cranking up the visual detail burns through the battery. Reducing power consumption generally means graphics that aren’t as sharp.
The answer, then, according to the company at least, is to move processing to the cloud as the next logical step.
Where cloud comes in
The cloud computing model of service-based application delivery and data storage, processing and analytics is widely agreed to be a logical step not just for edge computing, but for the majority of IT deployments. Where AT&T may be offering additional insight is in the expertise it can draw from its heritage in network transmission technologies.
The company says that in today’s networks, physical distances between users and data centers creates latency. As requests and responses travel hundreds or thousands of miles, users often notice the delay.
“With edge computing, we’ll install graphics processors and other computers in cell towers, small cells and other parts of our network that are never more than a few miles from our customers. This is what’s known as the edge of the network. In addition, low latency is being built into 5G from the get-go. The result: you will be able to run high-end applications in the cloud, and it will feel like it’s all happening right on your device,” said the company.
An Agile approach
Developers and other third parties will be invited to test and innovate at AT&T’s Palo Alto-based edge computing and, as with all R&D work, success is never guaranteed. But the company says its rapid innovation model (as in, Agile with a capital A) means it can move on quickly when an approach isn’t panning out and apply lessons learned to future projects.
“Our goal in this experiment is to find the right architecture, the right services and the right business value in this ecosystem,” said Igal Elbaz, head of AT&T Foundry. “It’s all about moving quickly and collaborating closely with third-party innovators and developers.”