As the big data sprawl continues to sweep across the enterprise, companies must remain competent in three component areas: storage, networking, and compute. Integrating enterprise AI storage solutions into business strategies is crucial for handling the massive data volumes efficiently and securely, leveraging advanced AI capabilities to optimize performance and ensure seamless operations.

With a streamlined platform in play, Pure Storage Inc.’s Digital Experience Business Unit is reshaping how businesses approach storage, with cutting-edge AI integrations and enhanced security measures poised to transform the experience into a seamless, service-oriented journey, according to Prakash Darji (pictured), general manager of the Digital Experience Business Unit at Pure Storage.

“When we think about a platform, you first need to ensure that you have a consistent operating environment,” Darji said. “A platform needs to ensure you get consistent improvement over time with a single unified infrastructure so you can support all types of workloads by deploying whatever you need. Most importantly, you need a single operating system. The top bun of the [platform] sandwich is the storage as a service, and the bottom bun is our Evergreen architecture. The meat of what you get is the Purity operating system with management at scale.”

Darji spoke with theCUBE Research Principal Analyst Christophe Bertrand during an AnalystANGLE segment on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed enterprise AI storage solutions and how Pure Storage’s DXBU is setting a new standard in storage solutions by integrating advanced AI capabilities, enhancing cyber resilience and redefining storage as a service.

Enterprise AI storage solutions: DXBU retools core platform for expanded capabilities

The DXBU’s core platform is a stratified whole comprising storage as a service at the top, the Evergreen architecture in the middle and Purity OS for storage at the final rung. Evergreen is designed to continuously improve IT models as it underpins incoming AI and security improvements.

Pure Storage has worked hard to integrate AI into the platform, according to Darji. Traditional AI models have evolved, with community models such as GPT-4 now training on millions of parameters to handle complex queries. The company is leveraging AI through a copilot, offering users an interactive, natural language interface to manage their storage environments. This copilot can provide security updates, operational assessments and actionable insights, making storage management more intuitive and efficient.

“You can use natural language. We’re making available this copilot in a technical preview to customers where it’ll go ahead and it’ll look at the customer’s environment and actually generate like, ‘These are the things that you need to focus on,'” Darji explained. “So, this is saying, ‘Across your administration, across your environment, these are the things that are fully implemented — these are the things that need improvement.’ Unlike it just being super open-ended, you want it to be conversational.”

As AI applications grow, so do the complexities of managing infrastructure. Pure Storage is addressing this by introducing the Evergreen//One AI service tier within DXBU. This new consumption model decouples performance and capacity, allowing businesses to scale efficiently and keep their GPU resources fully utilized, according to Darji.

“I give the people the analogy that it’s like the water bill in your home, where you have a pipe to your home. It’s a one-inch pipe, two-inch pipe, three-inch pipe, or four-inch pipe,” he said. “That guarantees the water capacity that goes into your home. That has to be the provisioned pipe in your home, and then you pay for the water usage that you use. We’ve designed this Evergreen//One for AI offering based on this paradigm where you have provision performance –, what is the throughput you need to keep your GPUs utilized at 100%?”

The Nvidia collaboration and strengthened cyber resilience

In partnership with Nvidia Corp., Pure Storage is enhancing its AI capabilities further. By certifying products with Nvidia’s DGX SuperPOD reference architecture, it aims to streamline the deployment of large language models on Ethernet networks, expanding beyond the traditional InfiniBand setups, according to Darji.

“A lot of customers have an ethernet network — we’ve been working with Nvidia to certify our products with ethernet into their SuperPOD reference architecture,” he said. “We’re announcing this concept of a secure application workspace where you can deploy staple containers for the data agent using Portworx, and that can create a vertical slice all the way down to a secure multi-tenant in the storage array.”

In today’s digital era, cyber resilience is paramount. Pure Storage is standardized with the NIST 2.0 framework, ensuring strong defense mechanisms against evolving cyber threats. Combined with the AI copilot, these capabilities are enhanced by proactive scans and operational security assessments, providing users with a comprehensive overview of potential vulnerabilities and actionable recommendations, Darji pointed out.

“We’ve aligned our protection philosophy based on the NIST 2.0 framework,” he said. “You can see that across identifying, protecting, detecting, responding, recovering and governing. We’ve had a set of capabilities, from safe mode snapshots to ransomware recovery. These are capabilities that are built into our platform, and we’ve consistently been improving with new capabilities, because as AI gets in the hands of the bad guys, your signature or variant attack patterns get more complicated.”

Here’s theCUBE’s complete AnalystANGLE segment with Darji:

Photo: SiliconANGLE