Pilot in Austin to Offer Early Look at Edge Computing at Scale
[ad_1]
The infrastructure needed to enable everything that 5G promises to make possible, from pervasive VR and AR to autonomous vehicles, goes well beyond new antennas on rooftops and cell towers. It’s going to need new fiber optic lines, computers, storage, networking equipment, sensors, cooling systems, power supplies, enclosures, physical security … you got the idea.
Keith Rutledge, general manager of Compass Datacenters EdgePoint business, compared the amount of infrastructure build-out to support the digital future that we had been promised for years to the electrification of the United States in the first half of the last century.
The cost of “cabling this country for autonomous systems and immense connectivity” will likely make it the most expensive infrastructure upgrade ever undertaken in the country, he told an audience at the Data Center World conference Monday in Orlando.
No single government or company can do this alone, Rutledge stressed. As with electrification or the railways, it will require the cooperation of numerous private and public organizations.
Compass, a Dallas-based data center developer and operator, is part of a pilot in Texas that is providing an early glimpse of what these public-private partnerships might look like in the future. Along with the Texas Military Department (a kind of oversized National Guard), numerous companies, including Compass, employ “a few dozen” PINNs or public infrastructure network nodes and a data center in the department’s Camp Mabry installation called Rutledge.
Compass EdgePoint provides the edge data center element of the project.
PINN, architecture of the Autonomy Institute’s Last-Mile Edge Computing Node
The program aims to develop a variety of “autonomous solutions” with a focus on autonomous vehicles. The project is being promoted by the edge cloud startup EDJX and the Autonomy Institute, a coalition with members from companies and government organizations that is committed to the standardization of 5G connectivity and the edge computing architecture around PINNs, one of which he has developed Concept.
It describes PINN as the “first unified open standard that integrates 5G wireless, edge computing, radar, lidar, improved GPUS and intelligent transport systems as a single unified system”. A PINN is basically a high-tech baton with built-in radio, camera and sensor devices, computing devices at the base, and LED lights for street lighting.
Institute for Autonomy
Autonomy Institute expects PINNs to one day enable “optimized traffic management, autonomous cars, industrial robotics, autonomous delivery, drones that respond to emergency calls, automated road and bridge inspection, smart city and national security applications,” the organization said in its announcement of the pilot project beginning this year.
A city the size of Austin (about 1 million people) would need a network of about 500 PINNs, Rutledge estimates. All 500 PINNs would also require four to five data centers to manage them. They could replace some cell towers, he pointed out, but they’re not meant to replace all of the towers.
The Autonomy Institute plans to expand its pilot PINN to other major US cities and eventually other countries.
Small data center networks for edge computing underway
There are of course others who build edge data centers and have different ideas about what this next generation of networks will look like. Austin-based Vapor IO has built clusters of small data centers (each about the size of a shipping container) on subways across the United States. The data centers are connected to cell towers, some are at the base of a tower and others not far from it. Denver-based EdgeMicro has also been busy deploying its version of edge computing facilities across the country.
All the different actors in this area seem to agree that there must be common framework conditions for many elements of the edge and 5G infrastructure and that several projects to develop these framework conditions are underway.
Vapor, for example, teamed up with a number of other vendors, including VMware and Dell, earlier this year to create the Open Grid Alliance, whose goal is to establish a formal architecture for the highly distributed network of the future. However, this alliance seems to focus more on the way edge and core data centers are connected than on the architecture of the end nodes through which data enters the backend network infrastructure, such as the Autonomy Institute’s PINNs.
[ad_2]