If you think Network Appliance's (NetApp) recent partnership with IBM was about getting the storage server vendor into the enterprise, Dan Warmenhoven wants you to think again. The company is already there, and enterprise customers are helping to fuel the company's extraordinary sales growth says NetApp's chief executive officer.
Like many high flying Silicon Valley companies, NetApp was hit hard by the dot-com bust, but the company's revenues are now growing at about 30 percent per year. The company, best known for selling NAS (network attached storage) storage servers that can easily be added onto the network, is looking for new areas to grow, and it clearly has one competitor in mind: EMC.
Warmenhoven, a former IBMer himself who has run the company since 1994, discusses the role of open source and virtualization software in Network Appliance's strategy and to share his thoughts on the blossoming relationship with IBM. Following is an edited transcript of that interview.
Though many of your competitors base their products on an open-source operating system, NetApp does not. What has been the effect of open source on your company?
It's been huge, not as a competitive threat, but as an opportunity. In the server world, open source really means Linux, and we believe that the adoption of Linux and Lintel style computers provides us with another one of those change opportunities... in the marketplace.
So we have been very heavily involved in the maturation of Linux as a server technology. We have been contributing to the development of the Linux/NFS (Network File System) infrastructure through funding of projects in the University of Michigan. We now have on our payroll the maintainer, (of the Linux NFS client) Trond Myklebust.
We're very involved in the Xen virtualization stuff because we think there's an opportunity for us there.
Tell me more about that.
I can't, but you get the idea. Open source comes in certain categories and rather than viewing it as a threat, we really view it as an opportunity.
So what is NetApp's plan when it comes to virtualization? EMC bought VMware and it seems that virtualization is going to become a much more important aspect of the storage business.
What VMware does is to take a large server and break it up into multiple small servers. That's really what they're good at. I think the next wave, though, is to take multiple small servers and aggregate them into a single virtual machine. And that, to us, is where the future of blade computing goes.
You can see the genesis for that kind of stuff in the... IBM Blade Center. But the real generic version is coming out of the early stage companies like Virtual Iron. Xen is probably going to do the same kind of thing in open source, we'll see.
The whole idea is that I want to treat CPUs (central processing units) as a commodity resource, and I want to be able to build a computing capability for a particular application that is truly flexible. If the application only runs for a couple of hours every day, I want to be able to deploy it when it runs, and I want to be able to deploy it to something else when it's done.
With EMC it's pretty clear that VMware is going to be the enabling technology for that. How are you going to achieve this at NetApp?
Open source and partnering. I don't have to own it to capitalize on that. EMC's acquisition of Legato was a great demonstration of that. They bought a 10 percent market share. My view is that it facilitated our partnering with the other 90 percent.
The day before that announcement, Veritas was in a position where they had a close partnership with Network Appliance, but they also had a close position with EMC, and they had to keep them relatively balanced. If they played favorites, they could wind up with a problem in the market sense.
As soon as the EMC announcement of the purchase of Legato went down, our relationship with Veritas went from good to great. There was no more ambiguity as to who the right partner was.
So are you thinking this virtualization technology is eventually going to be part of the operating system?
I'm actually thinking that VMware is not going to be the market leader in the next generation.
Another company you've developed closer ties with is IBM, which is now selling your products. Was the IBM deal really about getting NetApp into larger accounts?
No. It was geographic coverage, and other verticals we've never focused on. IBM is really focused on the public sector in state and local governments. We're big in federal, but we haven't done anything in state and local. IBM is really big in health care services. We've never tried to go crack health care services. Retail's another. We've never focused on retail. IBM's huge in retail.
It also has to do with market reach, in the sense that they're in a lot of countries we're not. I have no presence in Eastern Europe. I have two employees in Moscow. I think they're the only ones in the eastern European states. IBM's got a lot of business over there. South America -- we have a pretty good sized team in Brazil, but outside of that we have virtually no coverage.
This perception that we needed them to crack the enterprise is as far from the truth as you could possibly get. It's like saying we needed them to help our balance sheet or something. Look at the facts. We have penetrated the major accounts in a big way. My biggest account last year was Citigroup.
I'll put it in context. Last year our revenues were roughly US$1.6 billion. What do you think the IBM revenues are at my company this year? Maybe I should forecast forward. I think we're guiding the world to 30 percent year-over-year growth, roughly, so [next year] we're going to be $2.1 billion. What do you think IBM's revenues are in that forecast? How fast can they ramp up? I would be simply amazed if they're over $100 million. I'd be astonished. By the time IBM contributes anything, we're on our way to $4 billion. So this is not an "I'm dependent on IBM strategy."
You guys once used the Alpha processor, but have switched to Intel. What do you make of Apple's recent move to Intel?
I've always wondered why they stayed with PowerPC. I really did. PowerPC is not a bad microprocessor, but it carries a cost for it. They could get better price performance out of Intel because of the volumes. Either it hurt their margins or it hurt their prices, and Apple is always priced at a premium. I say, "Why do you persist in that?"
I guess the simple answer is that they had a way to differentiate themselves from Intel machines.
Boy, somebody's got to help me with that explanation. I've heard it before and I go,"Wow, that is the thinnest explanation I've ever heard." Let's see, the chip on the motherboard determines what the external characteristics are? You've got to help me here. I just don't get it.