Nvidia, Verizon using GPUs to run edge AI for real-time service enablement
As 5G continues to be refined through the 3GPP-led standardization process, consumer-facing benefits will morph into bespoke enterprise services that add low latency and high reliability to enhanced mobile broadband. In order to reap the benefits of real-time decision making supported by 5G, there needs to be concurrent investment in edge computing infrastructure in support of distributed workloads and the introduction of artificial intelligence for process automation.
Operators are already exploring how the combination of 5G, edge computing and artificial intelligence can support new services (and revenues) for things like facial recognition for public safety or access control. Similarly, enterprises are using the technologies to remotely control vehicles and add efficiencies to manufacturing.
Walmart, for instance, has set up an Intelligent Retail Lab that uses AI to analyze video streams and other sensor data to improve product inventory and availability, according to the company. “The team will use real-time information to explore efficiencies that will allow associates to know more precisely when to restock products, so items are available on shelves when they’re needed.”
While this doesn’t specifically incorporate 5G, the sensor data is fed into an edge data center where AI analytics are applied. This relies on NVIDIA’s EGX edge computing platform described by the company as enabling “IT to quickly and easily provision GPU servers” in a manner that “standardizes and automates the deployment of all necessary components for provisioning GPU-enabled Kubernetes clusters.”
Intelligent Retail Lab CEO Mike Hanrahan said in a statement that, with AI, “Customers can be confident about products being there, about the freshness of produce and meat. Technology enables us to understand so much more–in real time–about our business.”
Ahead of Mobile World Congress Barcelona next month, Nvidia is working to communicate the value of GPU computing at the edge to enterprise users like Walmart and also to operators. The company, last year, announced its carrier-focused Aerial software package that works with the EGX platform to “increase the spectral efficiency of their virtualized 5G radio access networks and offer new AI services for smart cities, smart factories, cloud gaming and more.”
On that vRAN point, Nvidia is working with network infrastructure vendor Ericsson. “The companies’ ultimate goal is to commercialize [vRAN] technologies to deliver radio networks with flexibility and shorter time to market for new services, such as augmented reality, virtual reality and gaming,” according to Ericsson.
Last year Verizon began testing edge GPUs for XR-type use cass. According to Verizon, internal teams put together a home-brewed “GPU-based orchestration system” that could “enable the development of scalable GPU cloud-based services.”
In a statement, Verizon said its team “developed a prototype using GPU slicing and management of virtualization that supports any GPU-based service and will increase the ability for multiple user-loads and tenants.” Tests focused on computer vision and a gaming service; in both cases, the new tech significantly increased the number of the concurrent users.
“Creating a scalable, low cost, edge-based GPU compute [capability]is the gateway to the production of inexpensive, highly powerful mobile devices,” said Nicki Palmer, chief product development officer. “This new innovation will lead to more cost effective and user friendly mobile mixed reality devices and open up a new world of possibilities for developers, consumers and enterprises.”
Verizon tested the combo of 5G and edge computing, also in Houston, using its 5G network and edge equipment to conduct AI-based facial recognition.
In an interview, Palmer noted, in conjunction with the GPU conversation, that Verizon in September acquired software, tech and other assets from XR specialist Jaunt, which was described as focusing “on the scalable creation and distribution of volumetric video of humans.”
“What we do with it, how we go to market…we haven’t detailed that,” she said. “This is just the beginning of a narrative.” When talking about 5G, you can’t just talk about radios, Palmer said. “You have to talk about what you’re going to do with it–help create the future.”
Cloud gaming, a partial focus of Verizon’s GPU work and a use case that has gotten a lot of attention from both device OEMs and their chipset vendors, but also from carriers in mature mobile markets like South Korea, could be in the works at Verizon. On Jan. 11, The Verge got an exclusive that Verizon is testing a video game streaming product that runs on Nvidia’s Shield set-top box.
“We haven’t detailed any plans here,” Palmer said, “but it’s easy to envision a mobile gaming service that has a slice of the network…that people would consume that has a certain latency profile and a certain bandwidth profile.”
