Intel becomes the latest hardware vendor to support Dell’s AI for Telecom services, with users able to run their AI workloads on Xeon CPUs and Gaudi GPUS, providing customers with more choices beyond existing options from Nvidia and AMD.
Dell’s service expansion also brings increased AI-powered solutions that will let operators implement AI to more efficiently manage their networks and develop revenue-generating services.
Subscribe today for free
A major area Dell is pushing with its latest telecom-focused update is around powering edge AI applications, such as visual inspection in manufacturing and smart transport.
Among the edge-focused solutions unveiled include an AI and Internet of Things (IoT)-powered security solution, created with Epic IO, that can be deployed to visualise sites like venues, providing improved visibility to help automate security operations
Also unveiled was an AI and machine learning-based framework solution made with Ecriothat that provides direct communication between operators and IoT devices, including capabilities like voice command and control of equipment like cameras, drones, and robots.
Dell also showcased an analytics tool created with TensorGo that uses video analytics of existing camera feeds at the edge to monitor foot traffic and crowd dynamics.
Andrew Vaz, VP of product management for Dell’s telecom system business told Capacity that Dell’s decision to focus on using AI at the edge with Intel was because it was an area that hadn’t previously gotten much attention from vendors.
“We're seeing more and more interest from telecom operators saying, ‘How do I offer something other than just connectivity to an enterprise edge?’ Vaz explained. “The Average revenue per user (ARPU) considerations are much better than just going after macro RAN like usual and offering connectivity services.”.
Powering the new edge use cases are several of Dell’s server solutions, with Vaz highlighting to Capacity that the modular XR8000 as seeing increased interest from operators to power their edge and edge AI use cases, including from hyperscalers.
“[The XR8000] has four sleds, a two modular power base which we can share the power across all four sleds in the back plane,” Vaz explained. “It is short in depth so it can fit in a constrained place and [a customer] can fit up to six GPUs 2u form factor or for some edge AI use cases you can use CPUs like the Intel Xeon.”
The Dell VP confirmed to Capacity that the popular server can also support a DPU offload card, providing customers with even greater flexibility in how they want to deploy the server.
"What we've found is that the use cases vary greatly, but the common requirements besides environmental factors are the need for compute resources, GPU capabilities, and offload capabilities,” Vaz said. “The ability to mix and match these components is very appealing, especially in a compact, ruggedised 2U form factor. This flexibility has resonated very well across various market segments."
Vaz suggested that if telecom operators were to use these AI solutopms to augment their edge work, it would offer a “massive win” for the industry.
“Telecom operators are embarking on a journey of monetisation not just typical connectivity but to use our platform to offer new services to the market,” he explained. “It’s popping in certain regions that are getting this, like with the Middle East becoming interesting.
“Other markets I didn’t expect include India when they’re building manufacturing sites, they don’t want to build things the same old way, they want to build with an advantage and they’re putting together these technologies to offer top-class new expansion.”
While he suggested it wouldn’t be a quick fix, Vaz said the opportunity edge AI solutions like those unveiled by Dell and Intel provide a “very tangible and massive monetary opportunity for operators who get involved.”
RELATED STORIES
Dell partners with Nvidia to accelerate AI adoption in telecoms
SK Telecom bets big on AI: Plans for AI 'superhighway' across Asia Pacific