A couple weeks ago, I was invited to attend the third Pure Accelerate show in San Francisco. I have been honored to be peripherally invited to watch this company as its growth into the industry has proven time and time again how well a company can spend time both supporting its existing customer base, and drive with an aggressive momentum the products forward.

I’ve seen how the initial product offering, the M series flash array had proven itself as an industry leader, and shown how the power of solid state can be both amazing in specific use cases, as well as a general storage environment, while still being very easy to manage. I then was privy to the launch of the FlashBlade architecture which leveraged unique, scalable architecture to increase both the storage footprint, as well as the processing power with each additional blade added to the environment. Then, the smart folks at pure launched the next generation of the Flash Array, the X Series array, which uses NVMe in lieu of traditional SAS or SATA solid state drives. Each of these steps taking into account some of the significant changes within the industry, most notably the NVMe architecture as a quantum step in solid state technology.

Amazingly to me, Pure has leveraged the same management platform for all the architectures, and have built within this platform, support for all storage types, intractability (is that a word?) amongst these, and given the scope of that software, still, all features, including replication, deduplication, and compression, among some of the less glitzy features are included with the array at no additional costs. I’ve spoken too of the amazing Evergreen program which, for the price of the maintenance contract keeps that storage environment current for not just the duration of traditional storage, but also offers processor upgrades non-disruptively on a roughly every three years interval. Thus, extending the life of the storage array to far beyond any promises made by other vendors.

I’ve spoken previously about the leveraging of NVMe-oF (or NVMe over Fabric) as an interconnect between Flash Arrays, and as a way to reduce the interconnect overhead into a sphere that was prior to this unheard of. As much as NVMe reduces bandwidth overhead from a processor to disc perspective, so does NVMe-oF reduce overhead between arrays. Thus giving a scalability functionality that few can match in terms of performance.

Also, last year, I was shown some of the applications being used that actually leverage some of that horsepower. For example, there was a facial recognition piece of software, for which I can certainly see many use-cases, though governmental, and security seem to me to be the most logical.

As I wondered what this year’s show would be promoting, I suspected that the application ecosystem that surrounds Pure would show more use-cases toward which the sheer power of these systems may be geared. Of course, I was not disappointed. Surely, there wouldn’t be yet another array, but hopefully, customers would now have use cases. And of course, I wasn’t disappointed. One particularly interesting use-case comes from the world of Biomedical Engineering, which has an app called Paige.AI which actively recognizes the growth of cancer from an Artificial Intelligence perspective. I anticipate some real progress in biomedical, seismic, petrochemical, high-frequency trading, and other areas far more typically associated with the High-Performance Computing (HPC) world, which will now be more commodity based, as a single sku line item can acquire a Flash Array or FlashBlade, whereas most HPC environments lean on parallel file systems like GPFS, and Panasas, or newer entries in the space like Weka.IO for those kinds of metrics.

Even the Evergreen program, mentioned above, and here has been expanded. The ES2 program, defined here takes the whole Capital Expense and turns it on its ear. As other vendors have taken the step toward leveraging the proximate data center to the major cloud providers, Pure is using this approach to “Rent” storage in these data centers on a pay only for the storage you use modality, and a true OPEX approach.

I find that the most compelling maturation this year, to restate, is that use-cases, and approach are changing the game for our friends at Pure Storage. Thank you for inviting me to this event, and allowing me to pontificate on what I see as very interesting changes beyond that of truly compelling hardware architectures.