This startup wants to be the Apple of the data center
Not every enterprise storage company looks to Apple for inspiration. But for Pure Storage, Apple’s combination of software and hardware skills is exactly what it wants to emulate.
Companies generate huge amounts of data that they slice and dice to help make decisions. But data-processing software is just one piece of the puzzle.
All that data has to live somewhere, and that’s why how companies store data is important. In the past, they used spinning-disc drives to house most of their data, but, over the past few years, flash storage technology has gained considerable attention for more efficiently retaining information.
A wave of startups focused on flash hardware quickly emerged to fill the void. But as the technology caught on, big legacy storage companies like EMC (EMC) and Western Digital (WDC) gobbled many of them up.
Pure Storage is among the flash storage survivors. John Colgrove, a storage veteran from Veritas Software, and John Hayes, a storage newbie, founded the company in 2009. The company boasts that its software smarts as well as its know-how on building hardware based on flash gives it an advantage. Pure’s clients include SurveyMonkey, Workday (WDAY), and ConocoPhillips.
The company is not yet profitable, but it says that its revenue tripled last year. In an interview with Fortune, Hayes declined to comment on plans for a possible IPO. The following transcript was edited for length and clarity:
Fortune: Why are people interested in flash?
Hayes: One of the reasons people love flash is because it’s so much faster, which means that you can quickly run queries. There are all these benefits that you get out of it, just simply being a lot faster. What we want to do is speed up all this batch technology.
What do you mean by “batch technology?”
Technology where you don’t care how long it takes to run analysis. You create backups overnight. You run analysis overnight. We’re taking those processes from the nighttime to the daytime.
There are a lot of things that happen in the data center that people only do at night. They only change their software or they only move hardware around. There’s a crew of people who come in at night and do it because they’re so afraid of changing stuff in the data center, they wouldn’t dare do it in the day while the business is running.
Why do we see so much emphasis on companies using flash right now?
Facebook (FB) used a lot of flash. Google (GOOG) used a lot of flash. How they did it was they made a custom architecture and they carefully built their own hardware and software. These businesses probably spend 20% on research and development whereas a typical corporation spends 2%. If you’re on a 2% budget, you’re not building your own custom hardware. You’re not writing all your software.
So you’re targeting companies that don’t have the expertise of Google and you want them to swap out all their storage hardware and software?
Yeah, just replace all the storage so it’s all fast. Because when you do that, you don’t have to replace the software anymore. All these enterprises that aren’t web companies, they’re on the two percent budget. All their applications now run faster [with flash].
What are some businesses that aren’t Web companies that are running flash storage and where is it helping them?
Let’s look at Nielsen (NLSN) They run analysis on data on a worldwide operation. Before they bought us, they could run that analysis once a day. After installing flash behind their application, they now run once an hour. You think about the difference between the returns from the analysis they do. There’s next day versus same day.
Storage has been the sick man of the data center for a long time. When you look at where disk is, it is more than a thousand times slower than everything else in the data center. It’s completely re-balancing how you think about access to data. Just getting that data is the bottleneck in any sort of analysis.
Why didn’t the legacy storage players tap into flash until recent years?
I think you never know what’s inside the strategic mind of another company. From the outside, it kind of looks like an innovator’s dilemma. Incumbent vendors have gotten used to the idea that they can buy innovation. They see a bunch of startups that they can pick the winner, and so they don’t feel like they’re behind. And you saw that when they bought flash startups, so they got paired up and some of them tend to market more successfully than others. The difference is that there aren’t many flash startups left.
What company’s business model do you want to emulate?
I would like to model us after Apple (AAPL). I think that it’s important that we’re committed to internal innovation. It does not have a vast, sprawling product line, and is extremely focused on customer value.
You want to be the Apple of storage.
Of the data center. How many projects are there in the data center? There’s storage, compute, networking. In theory, you can cover it in five products, right?
Is that where you want to branch out to?
Yes. The order is unclear.
You don’t hear enterprise companies referring to Apple a whole lot. What is it about that company that you can see applied to your business?
Well the comparison is against enterprise companies that have one of three paths. They make their one product and they stick to it, and sometimes it can grow to a certain size. Or they buy other companies and essentially become aggregators, or they become consultancies. None of those sound like fun businesses to be in. With consultancies you have to work for every dollar. If you’re just buying other companies, you’re making other people’s dreams come true.
So I guess that’s the other thing when I bring up Apple. There are very few companies that fuse hardware and software together. And I think that’s an important step along that path if we can develop new competence or new engineering competence and new sales competence. That’s a necessary thing to do to be able to create new products that you couldn’t create before.
What you’re saying seems to counter the trend of the Facebooks of the world buying commodity gear and loading custom software into it. Your company is packaging the hardware and software together, which has traditionally been what companies like Cisco have done. Why go back to that route?
Because if you’re selling commodity hardware, you’re basically saying that you don’t believe there’s scope for innovation in hardware, and I think that there is. There are a lot of craters for a company trying to innovate in hardware, and that’s because they’re trying to keep up with Intel or usually Intel, which is the big problem with hardware. Am I going to build something in hardware where if I did nothing for two years, I would get it for free from Intel? That’s the question you have to ask.
What about people who are afraid of vendor lock-in?
I don’t think there’s any such thing as vendor lock-in for storage. If you can get your data off, you’ve not lost it. And if we made a storage product where you couldn’t maintain data, it wouldn’t be very useful. Our worst case is actually the best case for the incumbent vendors.
Let’s presume every time your body changes, you will still be you. That’s how we think about data, and that’s how we think about how we construct hardware. Hardware wears out. Data is forever.
Subscribe to Data Sheet, Fortune’s daily newsletter on the business of technology.
For more on data centers, check out the following Fortune video: