BIG DATA
BIG DATA
BIG DATA
AI-ready infrastructure accelerates data science at scale
In a move to simplify artificial intelligence deployment within enterprise computing infrastructures, Pure Storage Inc. has teamed up with hardware maker Nvidia Corp. Using four Nvidia DGX-1 supercomputers, Pure adds its FlashBlade array to create AIRI, the latest tool for the AI-ready infrastructure.
“You pull compute and storage and networking all into this compact design so there is no bottleneck, that data lives close to compute and delivers the faster performance for your neural network training,” said Roy Kim (pictured), AI lead, director of products and solutions at Pure Storage. “All the enterprise needs to do is give [AIRI] to the data scientist and they get up and running.”
Kim spoke with Peter Burris (@plburris), host of theCUBE, SiliconANGLE Media’s mobile livestreaming studio, at theCUBE’s studio in Palo Alto, California, to discuss the technology behind the AIRI solution and its impact on the time needed to develop AI training models for data scientists. (* Disclosure below.)
Removing AI complexity
AIRI is designed to address the complexity surrounding an AI implementation process that can make splitting the atom look easy. The process often involves starting with a high-performance computing architecture like InfiniBand, powering it with a heavy-duty graphics processing unit and then blending in software, such as TensorFlow, for data programming. All of these moving parts can create a daunting task in some enterprises.
The jointly developed Pure Storage and Nvidia solution offers a preconfigured stack that removes a lot of the complexity and speeds up AI deployment. “Enterprises don’t need to worry about InfiniBand or Ethernet or GPUs or scale-out flash or TensorFlow,” Kim explained. “It just works.”
The new solution is engineered to provide four times the throughput, according to Kim. This means that where it might take a data scientist at least a month to start a workload and train neural networks, AIRI can reduce that process to a week.
“The scary thing about that is you have 12 tries a year to get it right, and that’s not something we want enterprises to suffer through,” Kim said. “Now they have 48 tries in a year.”
Watch the entire video interview with Kim below, and be sure to check out more of SiliconANGLE’s and theCUBE’s CUBE Conversations. (* Disclosure: Pure Storage Inc. sponsored this segment of theCUBE. Neither Pure Storage nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)
Photo: SiliconANGLE
Since you’re here …
… We’d like to tell you about our mission and how you can help us fulfill it. SiliconANGLE Media Inc.’s business model is based on the intrinsic value of the content, not advertising. Unlike many online publications, we don’t have a paywall or run banner advertising, because we want to keep our journalism open, without influence or the need to chase traffic.The journalism, reporting and commentary on SiliconANGLE — along with live, unscripted video from our Silicon Valley studio and globe-trotting video teams at theCUBE — take a lot of hard work, time and money. Keeping the quality high requires the support of sponsors who are aligned with our vision of ad-free journalism content.
If you like the reporting, video interviews and other ad-free content here, please take a moment to check out a sample of the video content supported by our sponsors, tweet your support, and keep coming back to SiliconANGLE.
