Accelerating to a Data-centric Architecture with Pure Storage

Having taken a few days break from the color orange, I’m ready to share my thoughts on Pure Accelerate. The show started with Pure Storage CEO Charlie Giancarlo talking about stools, with the three legs of a stool representing the core components of the data center: compute, networking, and storage. We were left with a new Pure vision anchored in storage, but focused on more than storage: a data-centric architecture that addresses the simplicity, agility, and performance requirements of the modern business. 

Data_Centric_Architecture

So what’s the data-centric architecture? An architecture that helps organizations consolidate and simplify the deployment and management of a globally-dispersed IT environment, respond to the business in real time, and scale and consume resources on demand with built-in autonomy, and that handles next-generation applications and workloads, and embraces the customers wants in terms of where those applications and workloads run, whether on-premises, in the cloud, or across multiple clouds. And while aspects of that vision are still aspirational for Pure, the focus on the future and where IT consumers will be in a year or two is exactly what got them to where they are today, which is constantly growing with a run rate over $1 billion dollars and all signs pointing up.

Throughout the main keynote we were given all the announcements. Key announcements included the new all-NVMe FlashArray//X family, $0 premium for mainstream pricing of NVMe, OpEx storage-as-a-service consumption with the new Evergreen Storage Service (ES2), a new FlashStack for modernizing Oracle data warehouses that can extend into AI and modern analytic workloads, and the AIRI mini.

While all the announcements deserve attention, I’m going to hone in on the AIRI mini announcement. AI is all the buzz and recently, Pure announced AIRI, which was focused on simplifying AI at scale. AIRI is a validated solution from Pure and Nvidia that can enable data scientists to be more productive by addressing design, deployment, and operational bottlenecks. And for organizations that are well on the AI path, this will serve as a solid infrastructure option. But what about those organizations that are just getting started? That aren’t ready to go deep into AI yet? That want to start small with a more cost-effective solution (compared with the AIRI) and scale as they get more comfortable? What about the organizations that have limited data scientists or no data scientists at all, but have citizen data scientists, data architects, or analytics experts?

Insert the AIRI mini, a consumable converged solution purpose built for AI that can meet the budgetary requirements of many organizations while delivering two petaFLOPS of performance (that’s two quadrillion floating-point operations per second). I’m excited to learn if/how Pure starts bringing a price/performance story to market, especially with the AIRI mini appealing to a wider audience due to its simplicity and price-point story. As AI becomes more pervasive in the market in terms of actual deployments, I’ll be interested to watch the traction the AIRI product line gains. And of the two offerings, while both will have their place in the market, my bet is on the AIRI mini to far surpass the more robust AIRI, enough so that Pure may want to rethink the name. Drop the “mini.” Aside from the physical size, there is nothing mini about the value organizations can potentially gain.

Watch this ESG On Location video from the event, in which my ESG colleagues and I break down some further insights from the event. 

Topics: Storage Data Platforms, Analytics, & AI ESG on Location