Q&A: Pure Storage CEO Dietzen Talks About NetApp/SolidFire, Flash Storage Future
Pure Storage And A Fast-Changing All-Flash Storage Business
The all-flash storage array business is on fire. The business has been driven first by startups that challenged large incumbent storage vendors with high-performance solutions, dropping flash storage prices so they compete with disk-based storage. But the incumbents -- including EMC, NetApp, HPE, Dell and others -- have responded with acquisitions of some startups or via their own technology, creating a very fluid environment.
Pure Storage CEO Scott Dietzen recently sat down with CRN for a discussion on NetApp's SolidFire acquisition, competing with the large incumbents like EMC, and continuing consolidation.
Turn the page for a whirlwind of insights.
Were you surprised by the news that NetApp plans to acquire SolidFire?
In general, not surprised. ... Our view of other upstarts in this space is that they haven't managed to get to anywhere near our scale. Nor were they growing as quickly. [In that situation] the notion of doing an IPO following ours would be tough. I don't know SolidFire's numbers, but assuming that speculation was right, that their business wasn't as large or growing as quickly, then M&A needs to be the exit they consider.
Flash is such a profoundly different media that the storage designs on the hardware side and the software side, on the cloud automation side, on the business model side, that all of them need to be hugely revised for the modern cloud-oriented data center. ... I think that is an endorsement from NetApp that they need new technology if they're going to play in this market.
This is similar to the moves that EMC made with its XtremIO and DSSD acquisitions, and we think [they validate] the fact that this new world calls for new storage design.
So what does this mean for the other startups in this field? Do you expect more acquisitions? Or companies to fall by the wayside? Is there room for startups?
My guess is the market is going to continue to consolidate. It's really hard. The intellectual property required to get flash right, in our view, is substantially harder than it was for mechanical disk, in the software and hardware designs, the cloud automation, and the radical change in the business model. ...
I think upstarts in this space need to be able to get to critical mass very quickly. I think right now, today, in our calculus, EMC and Pure Storage are the two solutions that are at critical mass. Obviously, others will have different viewpoints on what's a critical mass. But you've got to get to critical mass, and you've got to have triple-digit growth rates. You need to be growing faster than the aggregate all-flash market so you're building share. I think that's going to be a major challenge for the other players in this space.
You mentioned two companies with critical mass: EMC and Pure Storage, but not NetApp. In your view, does NetApp not have critical mass? If not, can it make it?
I said earlier that flash, solid-state media, is so different than mechanical disk that it just demands a different storage design. If you need new software, this multi-decades software that was crafted for mechanical disk is not going to be a fit for the solid-state and cloud-centric future. Then you're in a situation for any of the incumbents where they've got to get technology that is poised to support the cloud-oriented management and deployment model that customers want and that is capable of taking advantage of the solid-state flash media. ...
The challenge in front of a player like NetApp is that they've got this huge installed base that's on disk-centric technology, and they've gotta manage the same sort of transition that EMC is trying to manage in shifting customers from VMAX and VNX to XtremIO. I would argue that that's the hurdle in front of NetApp, and that it's a very tough thing to manage as a public company.
But NetApp has its EF line and AFF line of all-flash storage arrays. So they've taken steps in that direction, haven't they, especially now with the SolidFire acquisition?
The EF line and the FAS lines both have all-flash configurations. But they're both disk-centric designs that they've loaded flash on. In our view ... that doesn't get you into a position that lets you really take advantage of all of the benefits [of flash technology]. We don't think that 20-year-old technology, trying to retrofit that for the solid-state and cloud-centric future, we don't think that will be successful. Incumbent vendors are telling a different story, other than EMC. I would say EMC has been the most aggressive out there, leading the market with us and saying, "Yep, solid state and cloud require a new architecture." They've acquired not one, but two [architectures] to try and play in that space.
It'll be interesting to watch if NetApp starts talking more aggressively about that being the future, a more EMC-like vision of the future versus what they're currently articulating.
One could argue that, regardless of how EMC developed technology, the fact remains that it's an incumbent player that's very viable in the market. Would you agree?
I would. EMC has a significant amount of reach. I think they're going to have some pain with the Dell integration for the next year-plus. But we see them staying our strongest competitor over time.
Rather, I was highlighting [that] it's the rest of the storage players that are literally trying to retrofit 15- to 25-year-old technology for something that it was never designed to accommodate, which is flash memory, and this very simple deployment of scaling non-disruptive operations that the cloud demands. They're just not prepared for it.
What part of Pure Storage's business is through indirect channels?
I think there's a small part of the business, I think there are one or two customers that aren't channel-fulfilled. Virtually 100 percent of the business is fulfilled through the channel. And it's about half of [our business] that's channel-led. We identify channel-led business as [when] the partner brings the opportunity to us as opposed to us taking the opportunity to the partner.
Legacy storage vendors like NetApp and Dell argue that their architecture does work with flash, and they offer the kind of software and services that smaller companies focusing on this space like Pure Storage can't. How do you respond?
We're operating at a half-billion-dollar scale and growing dramatically. The reason we're performing this way in the market is [that] the technology we're delivering is highly, highly differentiated. Flash changes the way you design things. Each SSD can do about 64 things in parallel. [Incumbent vendors'] technology was designed for a mechanical disk that does one thing at a time. Flash can support variable I/O scheduling, [but] they use fixed-block architectures universally because it didn't matter in the disk era. ... We are two to five times better in data reduction. We're unique in generally embracing the most cost-effective flash, and doing so well ahead of the market. We changed the business model to support cloud where the hardware and software morph transparently underneath the data so [there's] no need to resell the same tech over and over again.
Do you have any examples of Pure Storage success in the cloud and Software-as-a-Service market?
So, six of the top 20 SaaS providers in the world are using Pure Storage. That includes Intuit, Workday, ServiceNow. Many of the top consumer Web properties are using Pure Storage in their infrastructure. LinkedIn and Shutterfly are doing so publicly, but there are many others that we haven't announced.
I can tell you that this is a supremely tough market for these legacy vendors to play in. Again, [their] products were designed for a mechanical device, not for solid-state memory.
I've heard said, "if it's flash, it's fast." Given that most of the vendors use pretty much the same flash suppliers, with some exceptions, could flash storage become more of a price exercise?
Our software allows us to take advantage of the most cost-effective flash. We were one of the first to embrace TLC. We do so in our mainline product. We don't need to do anything special with it, whereas other vendors, because of the way they treat flash like a mechanical disk, end up having to pay a premium and buying things like enterprise-grade flash because they just haven't done the re-engineering work around their software to be able to accommodate flash in the way it wants to be treated. ...
Things like snapshots and replication [must be built] differently in a flash world, because you build in technologies like deduplication underneath your data management services, because that's where it can be most efficiently done. Our competitors are trying to bolt them on after the fact.
Most of your legacy competitors are parts of larger companies with a server business, or have relationships with server companies. Does Pure Storage have any relationships with server-focused vendors to offer integrated solutions?
We're an equal opportunity disrupter, so we're working closely with all of the server vendors. You wouldn't know it from the marketing, but HP servers and Dell servers, they're very happy to embrace Pure Storage when Pure is what the customer wants.
I think we have the better go-to-market opportunities with Cisco and Lenovo, given that they don't have their own storage offerings. I think the Dell-EMC hookup probably makes Lenovo that much more friendly to Pure.
But I specifically want to highlight FlashStack because it's such a channel-friendly product. That's a collaborative [solution] with Cisco and with multiple different software suppliers. VMware was first, but SAP and Oracle have been drawn under the FlashStack umbrella as well. So that's an integrated product offering between Cisco UCS and Pure Storage in standard configurations. It's a come together in the partner community [solution].
That term "equal opportunity disrupter," did you just come up with that?
Ah, no, I confess, I've used that one in the past. But it is how we view our business.
Flash and cloud are such big changes to storage. The things you've been recounting about how the incumbents are saying they're fine is exactly what the tape vendors said, almost verbatim. And when Data Domain launched, [the tape vendors said], "Don't worry, we've got everything you need. You just need to stick with us. We've got the scale. We've got all the experience in backup. And we'll manage your transition to disk."
Boy, the industry sure didn't work out that way.
And you've got the biggest player in this space, EMC, endorsing our view that this is a wholesale technology change.
Does Pure Storage have a hyper-converged infrastructure strategy?
Rather than hyper-converged, we are on the converged bandwagon. That's what FlashStack is about. But let me add two thoughts on hyper-converged.
The current hyper-converged designs have a fixed ratio of compute and storage. That's not the way Web-scale works. [With] Web-scale, you need to be able to independently add compute and add storage in flexible ratios. It's all done incrementally. Inside Web-scale [environments], the compute and storage are actually separated into different nodes. ...
The other thing with hyper-converged products today is that they're designed around mechanical disk. ... So I do think there's an opportunity to build a much more flash- and cloud- and Web-scale-friendly hyper-converged architecture. But that doesn't exist today. That's why our energy is focused on converged, at least in tier-one [storage].
Pure Storage in 2015 said it would begin designing its own underlying hardware rather than depend on commodity hardware as a platform for running its storage software stack. How is that coming along?
The FlashArray//m series, which [went into general availability] in the third quarter, that is a form factor of our own design. It's designed to support our evergreen business model. And it gives us better costs. But it's all off-the-shelf silicon in that product. We're using standard Intel processors, standard DRAM, and standard flash we're getting from the top three flash suppliers which we have strategic relationships with.
So we're not doing our own silicon. But we are doing our own hardware because, with hardware designed for flash, we can get two to three times the density and power efficiency, and substantially better performance, than you can with these warmed-over disk designs, which, again, most of our competitors are using.
Will Pure Storage one day make its flash storage software stack available to other vendors to build flash storage using their own hardware?
The request has come up. Intel [processors] and DRAM are quite standardized in a way that makes them pretty uniform. Unfortunately, that's where it stops. The interconnects on the backplane tend to vary. And each generation of flash differs substantially from the prior generation in its operating characteristics. We do a lot of work to certify flash from different suppliers, and do all we can do that allows the product to get every bit of performance out of the flash that we're including. That's not work that you can just drop into any server with any SSD and get a reasonable result out of it.
Is Pure Storage in a recruitment mode for channel partners?
I would say, selectively. In general, our success is more about building deep relationships with really good partners than it is about getting lots and lots of partners in the door. We've focused on some really strong partners -- companies like ePlus, Datalink, WWT [Worldwide Technologies], Dimension Data, IAS [Integrated Archive Systems], ACS, SHI. That's where the preponderance of our energy goes, to those great strong, existing partner relationships.
So how is Pure Storage's market share? Do you have any numbers to share?
It's pretty tough to get. In the [most recent] quarter, we grew 167 percent year over year. That's higher than any of the estimates we've seen for the all-flash storage segment, which we believe is poised to replace all of tier-one storage over time. The consumer-grade flash we use, whether it's MLC or TLC, is now cheaper than a 15-K disk. And that doesn't count our five-times data reduction on top of it. Flash has really rendered that tier-one-performance disk obsolete.
We just need to continue to grow faster than the market to meet our aspiration to be the No. 1 storage vendor.
That 167 percent growth, that was in terms of revenue?
How about in terms of capacity?
You would be better served by [asking] someone like [research firm] Gartner who does industry supply checks across lots of different vendors. That's not something we have direct data on.
Has Pure Storage been talking to anyone about the possibility of being acquired? Are you on anybody's acquisition target list?
We couldn't comment on who might have us on their list. But I can tell you that over our lifetime, we've definitely been approached regarding M&A interest. It's not something that we've entertained to date.
We have a $24 billion TAM [total addressable market]. We've got the highest growth that we believe has ever been delivered in systems history. So that's not just in storage. We've beat all our storage predecessors. We've beaten the networking companies, Cisco in its heyday, Palo Alto Networks in its heyday. We've got the best team in storage. We've got this highly differentiated product, best customer satisfaction in the industry. That is the right recipe in business. That's all you need to build a No. 1 share position over time, and that's what we think we can do as an independent company.