Nvidia
Enterprises can begin to run trials of their AI projects using VMware vSphere with Tanzu together with Nvidia AI Enterprises software suite, as part of moves by both companies to further simplify AI development and application management.
By extending testing to vSphere with Tanzu, Nvidia boasts it will enable developers to run AI workloads on Kubernetes containers within their existing VMware environments. The software suite will run on mainstream Nvidia-certified systems, the company said, noting it would provide a complete software and hardware stack suitable for AI development.
“Nvidia has gone and invested in building all of the next-generation cloud application-level components, where you can now take the NGC libraries, which are container-based, and run those in a Kubernetes orchestrated VMware environment, so you’re getting the ability now to go and bridge the world of developers and infrastructure,” VMware cloud infrastructure business group marketing VP Lee Caswell told media.
The move comes off the back of VMware announcing Nvidia AI Enterprise in March. At the time, Nvidia AI workloads were only able to run on vSphere 7 certified virtualisation platform. By August, Nvidia announced the general availability of Nvidia AI Enterprise to allow companies to virtualise AI workloads on mainstream servers running VMware vSphere.
“Our engineering teams are aligned and the real goal around this is to really make sure that we deliver customers, fully functional AI platform as well as from a performance perspective so adding to Kubernetes grid is a fantastic capability,” Nvidia enterprise product management VP John Fanelli said.
Additionally, Nvidia announced it will Lenovo to its early access program to allow customers who currently use Lenovo systems to participate in VMware’s Project Monterey.
Project Monterey, announced during last year’s VMworld 2020 conference, was considered by VMware as focused on redefining hybrid cloud architecture for the data centre, cloud, and edge.
As part of the early access program, customers will be able to remotely access Project Monterey running over Dell EMC PowerEdge R750 and equipped with Nvidia’s Bluefield-2 DPU. It will be hosted within Nvidia’s environment.
Fanelli believes the program could be a potential driver for early security uses cases.
“The focus of this is around the ease of access for our customers to try this out,” Fanelli said.
“We’re hosting this initially at an environment at Nvidia, so that we can provide access to customers — as you know, there’s all kinds of issues around supply chain and getting server set up — so we really wanted to simplify it.”
