One of the few private vendors left in the all-flash storage market, SolidFire, is taking a different route than...
its better known -- and newly public -- rival, Pure Storage Inc.
Both vendors began shipping arrays in 2011, but Pure quickly amassed more than $530 million in funding, ramped up sales and became a public company in October. SolidFire took a more measured approach, picked up $150 million in funding, and set out to build a base of large enterprise and cloud-provider customers. Pure racked up nearly three times the revenue as SolidFire in 2014, according to research firm Gartner Inc., but also incurred heavy losses.
SolidFire founder and CEO Dave Wright said his strategy is to steadily grow its base and hit profitability quicker than Pure. SearchSolidStateStorage recently spoke with Wright about SolidFire's plans and the all-flash storage market.
How has the conversation around flash changed since you started shipping arrays in 2011?
Wright: The perception of flash in the enterprise is changing, and customers realize it can cover a significant amount of their workloads. That's been part of our message from the beginning, that flash is not a niche solution. With the combination of falling price and data reduction technologies, you can apply flash across a huge chunk of your data center footprint. A lot of enterprises are starting to realize that, having fought back against the FUD from the big vendors. They’ve seen through the claims that flash is a unicorn that you keep in the corner. You can use it all over the place, and it has tremendous benefit beyond just performance. Its power, space, cooling and reliability are all greatly superior to disk.
Is it enough for a systems vendor to only sell all-flash arrays, or will SolidFire have to broaden its product portfolio?
Wright: I'm not one to believe that flash is going to completely replace disk in storage systems in the short term. There will continue to be a role for disk. It will more and more fall into the worlds of backup and archiving, and cold storage and content repository use cases, where it's storing large amounts of data for a long time without much activity.
But, in terms of what you need to be a successful storage company, I think the opportunity in flash is pretty significant. If you think about the continuum of applications, and what's on disk today and what's on flash today, the line between disk and flash is only going to shift in one direction. More and more workloads will be more cost-effective to put on flash; more and more workloads are going to be displaced from disk to flash. Disk is going to get squeezed further and further into the corner of only the coldest and slowest workloads of data.
Don't get me wrong, that's still going to be a tremendous amount of data, because capacity-wise, a lot of data is cold, dumb data that just sits there and doesn't do anything. But everything else will move to flash in the next couple of years. That's going to lead to a rapid expansion of the market opportunity for flash. It makes far more sense to focus on expanding the roles and use cases in the data center that an all-flash system can handle, rather than try to build a disk-storage system, which is ultimately a shrinking market. A much faster growth opportunity is in flash.
But, as we've seen with Dell-EMC, large vendors are amassing all the pieces of the stack. Where does that leave the smaller guys?
Wright: We do that through partnerships. Through Agile Infrastructure [SolidFire's reference architecture program], we have partnerships with Dell and Cisco. Customers like having all the pieces of the box coming from one vendor, but they also like having the flexibility to choose who to work with. I don't think you have to do it all yourself.
With all legacy vendors now having all-flash systems, how many flash-only vendors can survive? Do you have to establish a foothold quickly?
Wright: A number of all-flash array vendors have already consolidated. We've seen Whiptail, Skyera and a few other smaller vendors disappear, and there really aren't many all-flash startups remaining. You can look at it as a land grab, where there's a race to get as many customers as you can as quickly as possible. And that, ultimately, is your beachhead. But the huge vendors already have such a huge head start there, and that's not really a viable strategy. They have tens of thousands of customers; it's impossible to outrun those guys to simply get more customers. Our strategy has always been to simply be a flash storage system and focus on a differentiated set of capabilities beyond the flash that other vendors aren't able to do.
We can go to a customer and talk to them about things we can enable in their data center that they can't get elsewhere. We can talk about things like a shared-nothing, scaled-out architecture that gives them ultimate storage agility and flexibility, guaranteed quality of service and performance, and really rich automation capabilities. All of these things aren't directly related to flash at all, but we built all of them into our architecture. Our conversation with the customer isn't about flash versus disk; it's a conversation about all of the other things they're trying to do to modernize their data center. That, ultimately, leads to a different conversation than selling flash.
SolidFire and Pure Storage started around the same time. Now, Pure is a public company and third in all-flash storage market share, while you're in single digits. How come they're so far ahead of you in sales?
Dave Wrightfounder and CEO of SolidFire
Wright: We've raised a significant amount of money -- upwards of $150 million -- but it's very different from how they've raised funding, and how quickly they've hired and grown their field sales. They have a good base of customers, but they're spending a lot of money to acquire those customers and they're burning a lot of money. The market we're going after is more focused on large enterprises, service providers, telcos and more complex data centers. That market has a different selling motion and sales cycle than Pure's. We've built our company around a much more strategic approach to selling to that market.
It sounds like you don't go head-to-head much against Pure.
Wright: In most of the larger opportunities, we compete with the larger vendors -- EMC, NetApp, [Hewlett Packard Enterprise] 3PAR. And often, it's their hybrid systems, although increasingly they are pushing all-flash arrays.
Do the large vendors push hybrids on price?
Wright: Vendors are trying to prey on customers' preconceptions about the price of flash and convince them a hybrid solution will be better for them, even if it's not really that much less expensive. Quite frankly, it's a lot more profitable for these [large storage] companies to sell hybrid systems than their all-flash array. In most cases, their all-flash arrays haven't been architected from the ground up around commodity hardware and consumer MLC flash the way our system has, and the way Pure's system has. So, to price competitively, they have to discount heavily.
They're pushing all-flash systems more now because that's what customers are looking for, but they still have hybrids in the game.
Are you finding customers who only want all-flash arrays and won't consider hybrids?
Wright: Absolutely. But it's not typically first-time buyers who want all-flash systems. It's customers who have bought flash for point workloads or for performance acceleration, and they've seen the benefits of flash and the price of flash come down. It's a tipping point when they think about investing in a new storage system they'll be dealing with for next three to five years. They're just done with disk. It doesn't mean they'll never buy another disk-based storage system, but when it comes to primary storage use cases, they're not going to mess with disk. It's not worth the time, hassle and headaches, and not worth tradeoffs of hybrid storage system. When they think of their storage system today, and where they're going to be three years down the road, they realize they're putting a boat anchor on their leg if they buy hybrid system.
How low can the cost per gigabyte reasonably go for flash?
Wright: We've seen the cost of flash fall 20% to 30% per year every year in the five years since we've been around. With advances in flash technology like 3D NAND, I think you'll continue to see the price drop at that rate for the next few years. Now, what's the actual cost per gigabyte? A lot of vendors play a lot of games assuming data efficiencies, discount levels, configurations. There are complexities and nuances in enterprise storage systems.
Speaking of 3D NAND, several vendors have started shipping it in flash arrays. When will SolidFire offer it?
Wright: We've been shipping it for a while. We haven't made any big public announcements. Others have disclosed it, but we don't consider it all that important. Our systems adopt the latest and greatest technologies, the price comes down and the architecture of our system allows us to mix and match.
How will the Dell-EMC deal merging two large competitors affect SolidFire?
Wright: It will create uncertainty [for Dell and EMC]. It will effectively neuter one of the things that EMC has used against startups -- that you can't trust what's going on with those guys, their product may go away. Now, EMC's in that position. They got bought out and they're about to have a lot of turnover. Now, there's a stronger argument for taking a look at fast-growing companies. EMC says storage startups are gnats buzzing round, but a swarm of mosquitoes can be annoying.
How to meet capacity needs with all-flash
Available features in the all-flash storage market
All-flash storage may be the right choice for you