Olivier Le Moal - stock.adobe.co
Pure Storage CEO Charles Giancarlo said his company still has plenty of work to do to help customers get to the next stage of digital transformation and does not rule out more acquisitions to get there.
Giancarlo took time out in the middle of the monthlong Pure//Accelerate Digital 2021 conference to speak with TechTarget about the company's strategic direction in the wake of the COVID-19 pandemic, its 2020 Portworx acquisition to focus on storage for container-based environments, the latest trends with subscription-based services and cloud options, and his take on flash drive manufacturing and 3D XPoint memory.
More than a year has passed since the pandemic started. What are the biggest takeaways for you as a CEO?
Charles Giancarlo: Shortly after it, I got COVID, so it certainly gave me a lot to think about. I suppose we're not quite out of it yet. But one of the most amazing things to me is the resiliency of the economy and society, given how serious not just the pandemic itself but also the response to the pandemic has been [with] the work-from-home environment, closing of businesses and so forth. Not only did business bounce back, and employment [is] bouncing back, but also the U.S. economy has been resilient.
One of the more humorous things has been watching the Ferris wheel of opinion -- everything from 'Business will never be the same, and everyone will be working from home from now on,' to 'We're going back to work in some form,' and now 'We're going to be entirely back in the workplace' -- and all of that, in some cases, from exactly the same prognosticators.
Lastly, while there was a lot of rhetoric around a focus on digital transformation, now everyone understands that it's actually not just a 'nice to have' but absolutely critical whether it's a pandemic or other national emergencies. Having a business based to a large extent on digital is what allows corporations to survive.
What is your sense of how far along your customers are on the journey to digital transformation?
Giancarlo: There's a full spectrum. About 30% of our sales are into SaaS companies. They're the ones that have been driving digital transformation and are largely, fully digital businesses, so they've fared quite well. Then you have companies on the other end of the spectrum, where it was to a large extent business as usual and bricks-and-mortar. They've had a much more difficult time, and they are much earlier in the phase. It's hard to speak about the average customer, but I would say the vast majority are still in the earlier stages.
Do you think Pure Storage has all the necessary pieces for customers focused on digital transformation?
Giancarlo: I wish that were true. If that were the case, I would be spending less on R&D. We still have a lot to do to help our customers get to that next stage. We're in one of the oldest of the tech industries. Data storage dates well back into the '50s, and in many ways, data storage is the last of the major infrastructure businesses to migrate into a more virtualized, API- and code-centric world. Storage has been largely a manual process. It's the last one to get to full solid state. It has been based on spinning disks -- mechanical storage -- and tape, also mechanical, for the longest period of time. It wasn't until the advent of both flash and high-speed networks that allow storage to be connected through a standard communication system -- rather than DAS or Fibre Channel -- that now we're finally entering the modern age where storage can be modernized, virtualized and enable developers to get access to it solely through code. That will allow our customers to finally get fully into digital transformation. Now we operate very much at the infrastructure level. But until the infrastructure is automated, the customers can't enable their developers to be fully automated.
Do you have the know-how in-house to cater to customers focusing on DevOps and digital transformation? Or will Pure Storage need to acquire more companies, as IBM did with Red Hat?
Giancarlo: We largely have that capability in-house, but we probably will go to the outside both for technology and talent. I'm not eschewing acquisitions at all. We acquired Portworx. It brought some great talent and capability into the company.
Charles GiancarloCEO, Pure Storage
Is Portworx the main piece of the equation for Pure to enable digital transformation? Or are other components equally as important?
Giancarlo: We have the foundation already. Evergreen is a foundational element in the sense that we can do nondisruptive upgrades of our technology. If you think about any SaaS service, or even the hyperscaler services, they're upgrading their systems all the time nondisruptively to the customer. So, the customer just sees better and better performance and capability every week, every month. We can do the same with our products based on our operating systems and the design of our systems overall. So that's a very important component. What Portworx brought in was the ability to do the same in a container space, which is an important new area for developers.
We've also enabled our IT customers to manage their environment through our cloud management system, Pure, and through APIs. The next step is enabling their developers to be able to define the storage that they need through APIs and through a web presence. We now have that on the Portworx side for containers. We're developing the same capability for traditional applications [that deploy on] virtual machines or bare metal, like databases, Oracle and homegrown applications.
What additional levels of integration can customers expect between the Portworx software and Pure Storage products?
Giancarlo: Portworx really consists of two things. One is the software-defined storage, and the other is the Kubernetes layer that allows for the orchestration of data management, as well as applications above the data layer, to get access to that data. That Kubernetes layer is going to be extended across all Pure products. That's one area of integration. The other area is that we are going to continue the software-defined storage layer for containers of Portworx. We think of that software-defined layer as an excellent layer for development, because it starts off very small. It scales as the development system gets larger. It can even go into production. When customers want large-scale production systems, we're giving them the ability nondisruptively to swap in a FlashArray or a FlashBlade underneath to either supplement or to replace the software-defined storage system with a system that is even more reliable, more robust and more feature-rich.
Do you see Portworx software-defined storage as a major revenue producer?
Giancarlo: We spent a lot of money on that acquisition. We're expecting revenue. Really what I want to see first is the Portworx capabilities in a lot of customers' hands, because it starts off as a development system. Container-based systems right now in production are relatively small. But the number of development projects starting off as containers is very large. So I just want to see Portworx in use because that's the leading indicator for revenue down the road. Portworx revenue was exceeding its pre-acquisition plan by 20% over the last couple of quarters, and I'm expecting good growth out of that product both on a usage and a revenue base for years to come.
Do you have any sense of how many Portworx customers are also Pure Storage customers, and how many use other vendors' storage?
Giancarlo: That was a big part of our diligence beforehand. Roughly 50% of their customers were customers of Pure. [Those Pure-Portworx customers] tend to be large enterprises that have very significant consumer businesses. About 50% are IT teams buying it for their developers, and about 50% are people outside of IT buying it directly. When we interviewed the developers buying it for themselves, they said, 'We're doing it now for ease, but eventually, we want IT to take care of this. We don't want to be taking care of storage and data management. We just want to focus on our application.'
Giancarlo: You have to scratch the surface to understand what they're really defining for their customer base. In our case with Pure as-a-Service, we negotiate a service-level agreement with our customer. That service can be delivered on prem. It can be delivered in the cloud. We deliver it, manage it, take care of it, upgrade it -- all done transparently to the customer. And over time, we're delivering a set of APIs to our customers' developers, so that they can just call upon the data management services that they need solely through code.
How popular is Pure as-a-Service with your customers compared to buying equipment outright?
Giancarlo: We haven't given exact percentages on how much is PaaS and how much is a capital sale, but the PaaS portion is growing much faster than the business as a whole. All our subscription businesses together -- which is the combination of Evergreen and Pure as-a-Service -- grew 30% last year. Our total business as a company only grew single digits, largely because of COVID.
What do you expect to see in the future?
Giancarlo: I suspect that in less than five years, our subscription business will exceed the capital portion of the business. But in terms of new buyers, I suspect that it may take a little longer than that before half the buyers are starting from scratch as-a-service versus capital. It'll be largely because of the preference of the customers themselves based on how they manage their P&L and balance sheet.
Why do customers need Pure Storage in the cloud? Couldn't they just go to AWS, Microsoft Azure, Google or whoever? They've been developing more of a hybrid story with their on-premises offerings.
Giancarlo: We have Cloud Block Store and Portworx both operating on AWS and Azure. Additionally, Portworx also works on Google Anthos. We're the only vendor that has been certified to operate on [AWS] Outposts with both FlashBlade and FlashArray. We add value with our Cloud Block Store on both AWS and Azure -- everything from lower costs to more features and functionality than would exist with a bare-bones service on those offerings.
But more importantly, we support traditional application environments with all the features that those environments expect from block storage, which typically you don't get in the hyperscalers that are designed much more for container-based environments. And we make the lift-and-shift migration from traditional applications to the hyperscale much easier. Of course, most of the application work going on with the hyperscalers is with so called cloud-native workloads. We help there as well with Portworx, which supports cloud-native workloads, but provide all the data services that customers don't necessarily want to develop on their own.
AWS has ambitious plans in the area of storage. Down the road, do you see them as competition, or do you envision traditional vendors such as Dell and HPE as your chief rivals?
Giancarlo: I'd probably be blind if I didn't say that Amazon could become a dangerous competitor. But I can't think of one time when we were competing directly with Amazon. We're competing with Dell, HPE, Hitachi, NetApp each and every day. Way back when, I once had Microsoft basically telling me that they were going to compete with everything we do: Why don't we give up now? So, the same is true with Amazon.
Pure still makes its own flash drives, and many of your competitors don't. Do you still see significant advantage in manufacturing them?
Giancarlo: Yes, 100%. If we didn't see a significant advantage in making our own hardware, we wouldn't do it. Remember SSDs are designed specifically to look like magnetic hard drives so their customers did not have to change their software in order to use them. We don't make disk drives. We make cards that have flash so that our software can talk directly to the flash and make the best use of it. Because of that, we have almost twice the longevity. We make nondisruptive upgrades easier. And we're probably 10 times more reliable than much of our competition.
Are the reliability and longevity also true for the consumer-grade QLC flash that you use?
Giancarlo: We don't know yet. We only started producing it last year. So, we don't have full data on it yet. We're likely to have slightly lower longevity. But, the jury's out on that. We don't know fully until we have more years of experience.
Do you think flash is enough for most customers? Or do you think you'll need new memory technology such as 3D XPoint?
Giancarlo: We've looked at it. We've thought about it. We experiment with a lot of different things. We think flash is going to be not just good enough, but it's also going to be the best for quite some years to come.
What did you think of Micron discontinuing 3D XPoint development?
Giancarlo: We were a little surprised, but I think it's a recognition that it doesn't really fill a niche that other technologies can't fulfill with a larger volume.
Editor's note: This interview was edited for the purpose of length and clarity.
Carol Sliwa is a TechTarget senior writer covering storage arrays and drives, flash and memory technologies, and enterprise architecture.