August 25, 2023
Simplifying AI with Pure Storage
To keep up with emerging technologies such as artificial intelligence, organizations need infrastructure that reduces complexity.
With artificial intelligence dominating headlines (and, according to some, largely responsible for this year’s stock market rally), customers are asking us about the technology — a lot. But when they ask where they should be making investments to accelerate their own AI programs, they’re often unaware that successful initiatives require both strategy and infrastructure.
While AI algorithms and software are crucial, having the right infrastructure in place is the critical foundation that makes AI implementation truly possible. Many overlook the importance of modernizing their data pipelines, moving to the cloud, and leveraging graphics processing units and specialized AI hardware. But these capabilities allow companies to store massive data sets, access scalable computing and run complex neural networks — the backbone of enterprise AI. So, while AI itself grabs attention, for long-term success, strengthening infrastructure should be an equal priority. With a robust foundation in place, companies can then layer on advanced AI and see faster time-to-value.
It’s important for organizations to work with infrastructure vendors that understand how urgent an organization’s AI initiatives are and that can take steps to support them. One example I often bring up is Pure Storage.
I recently had the chance to speak with Justin Emerson, principal product manager and technical evangelist at Pure Storage, and I walked away from the conversation impressed by how much thought and care the company is clearly putting into its customers’ AI programs. “There’s still a lot of confusion around AI, but people feel like if they don't do something, they’re going to be left behind,” he told me. “That’s why we’re trying to simplify things. There’s enough complexity in this space already, and the last thing we want to do is waste our customers’ time.”
Here are three ways that Pure Storage is helping its customers make their AI visions a reality.
In our conversation, Emerson pointed out that, as a younger vendor, Pure Storage has come up in the age of flash arrays and automated telemetry. Also, he noted, the company has grown organically, rather than through mergers and acquisitions. The result: straightforward storage infrastructure with intuitive management capabilities.
In fact, Emerson told me, Pure Storage has a goal of making the administrative tasks for its products only 20 percent as time-consuming as those of other solutions in the marketplace. “In other words,” he said, “you need one-fifth of the people to manage our systems compared with complex legacy storage systems.”
Emerson noted that many AI initiatives start small, on a single server or even a laptop in some cases. But, as these projects show promise and begin to scale, they often require significant resources — not only computational power but also massive amounts of power and cooling. (One recent estimate put ChatGPT’s daily power consumption at more than 260 megawatt-hours.)
These energy demands can drive up costs for organizations and make it challenging for them to meet their goals for corporate sustainability. Energy efficiency is a hallmark of Pure Storage infrastructure. “Our direct flash management system enables us to deliver systems that can consume one-fifth of the power of an equivalent all-flash system — or even less compared with spinning disk systems,” Emerson said.
Next-Gen AI Infrastructure
Pure Storage describes its AIRI//S infrastructure, powered by NVIDIA, as the industry’s “first complete, AI-ready infrastructure.” The infrastructure can be set up, deployed and managed quickly as an end-to-end AI pipeline solution, can seamlessly scale, and includes software and hardware updates that adapt to changing needs without disrupting operations.
“We’re focused on how we can deliver outcomes to customers that simplify the entire data pipeline,” Emerson told me. “We want our customers to be able to spend their time on their AI research and development, rather than becoming bogged down in managing infrastructure.”
Story by Rich Pushard, who serves as Product Portfolio Manager at CDW for the Higher Education vertical, applying his background in technology and innovation to higher education.