LAS VEGAS –– Organizations are finding real use cases for augmented reality, but IT infrastructure admins must first implement technologies that can help support it.
Edge computing, which processes data close to its source to reduce latency, is one way to support augmented reality (AR). IT can implement edge computing architecture by deploying edge servers, adopting micro data centers, or both. Here at Gartner’s IT infrastructure, operations and cloud strategies conference, admins discussed how AR and edge computing work together.
“A lot of [AR] has to do with visual data processing,” said Christopher Hadley, a compute architect at Georgia-Pacific, a paper manufacturer based in Atlanta. “You don’t want to have a lag on that sort of stuff, so putting the compute closer to where your [devices] are, the better.”
AR use cases
Manufacturing and retail are two major verticals that could benefit from AR. Georgia–Pacific is working on edge and AR initiatives within its manufacturing facilities, many of which do not allow humans in them, Hadley said.
With AR, Georgia-Pacific can use lesser-skilled workers to perform maintenance tasks in the facilities that do allow humans, rather than having an engineer on site at all times. Instead, an engineer works remotely using an augmented reality platform to instruct that worker how to perform more complicated tasks.
One consumer goods manufacturer also considered AR capabilities within manufacturing plants, according to a director of computing services at the company, who requested anonymity because they were not authorized to speak to the media. The organization has 50 manufacturing sites, many of which aren’t located in areas where they can get a seasoned engineer on site.
“If we’re running an [assembly] line and that line isn’t running as we expect, it’s a lot easier to have somebody with AR go through and be able to fix that in real time,” the director said.
There are also use cases for AR within the data center itself, but that is more of a concept than a reality today, said Jeffrey Hewitt, a research vice president at Gartner. Data center infrastructure management (DCIM) tools already provide a visualization component to help admins facilities management capabilities such as cable management, server identification and airflow management. DCIM vendors will likely tap into AR capabilities in the future, Hewitt said.
Edge computing architecture and AR options
There are a variety of challenges around the implementation of edge computing and AR, because the market is still in its incubatory phase. That means that IT needs to find creative methods for deployment.
“If you’re going to have an AR capability, the factor to make that more challenging, or even keep it from working, would be latency,” Hewitt said. “AR has to happen rapidly. It can’t have delays.”
Georgia-Pacific deploys micro data centers and is looking at hardened thick edge devices — essentially virtual machines in a micro data center’s virtual server farms — to achieve an edge computing infrastructure, Hadley said.
The process for choosing edge devices is based on an organization’s particular use case. Georgia-Pacific, for example, wants to put cameras and sensors in its outdoor wood yards to gauge information such as water levels and the volume of raw material. Outdoor wood yards, as well as the company’s manufacturing plants that face high temperatures, humidity and vibration, are considered harsh environments.
“We’re looking at some … very hardened edge servers,” Hadley said. “They’re completely self-contained, waterproof and don’t require fans or a lot of power. We want to get a bunch of those and spread them out to where the loads are.”
Vendors such as Dell EMC, Logic Supply and HPE offer these types of products in the form of ruggedized servers powered by Intel Xeon.
The consumer goods manufacturer deploys micro data centers in a DIY configuration at its manufacturing plants, and it is considering using Microsoft HoloLens for AR, the director of computing services said.
The organization uses Infosys, a managed services provider, and wanted the plants to have the same look and feel as the main data center. Like the primary data center, the micro data centers are VMware environments that run Cisco and NetApps’ FlexPod converged infrastructure.
“We’ve taken what we run in our data centers… and made a miniature size of that for our manufacturing sites,” he said.
“It’s technologies that [our managed services provider] already knows. We needed to ensure that they would support it.”