March 23, 1999

TMF Interview With
Network Appliance CEO
Dan Warmenhoven

With Dale Wettlaufer (TMF Ralegh)
Conducted February 25, 1999

About six months ago, I heard about Network Appliance (Nasdaq: NTAP), a data storage and retrieval company whose stock has quietly gained nearly 100% per year, compounded annually, since coming public in 1995. This is one of those stealth companies, like printer manufacturer Lexmark, that has compiled excellent returns to shareholders without wide notice outside the people who cover them and, of course, their shareholders.

When a company has come out of seemingly nowhere for the generalist investor, a common reaction is "Well, I've just missed the boat." While I can't express an opinion on that matter for NetApp, I will say that "missing the boat" is about the last concern that makes sense in this sort of situation. First, you want to understand such a business. That takes some basic questions before asking about the differences in SANs, NAS products, and the like.

For the technically oriented reader, this interview may not tell you that much. I wanted to ask some basic questions to help investors who are new to the company. The following interview with Network Appliance President and CEO Dan Warmenhoven is offered in that vein. To learn more about Network Appliance check out the company shapshot.

Dale: What's your sales proposition, Dan?

Warmenhoven: This is going to be faster than what you're currently using; it's simpler; it'll reduce your operations complexity because of its simplicity; it's more reliable and more available than probably anything else you can buy. It's more cost effective in terms of dollars per gig or dollars per user support. And that's a pretty compelling value proposition to put in front of customer. Most of the time that list of benefits gets their attention.

Dale: When you have a salesman go out to a customer, what exactly are the metrics that you base the presentation on?

Warmenhoven: It varies quite dramatically based on what the potential application is and what the organization's customer is. If you are selling into a technical community, and in particular let's say software development environments or ECAD environments, they are generally very, very orientated towards raw performance. And that translates into your productivity improvements -- faster build times for software products or faster chip simulations, which gives more runs per day, if you will.

If you're talking, however, to a CIO and let's say you're in a Windows environment, which is probably true, it typically gets down to simplifying the operating environment so you can kind of cap the growth of staff, and it also improves the overall availability of data to the those end-users.

If you shift to database applications, it is very much orientated towards data availability, data recovery, and simplified data management.

And then if you go into the caching arena, you know, the Internet and intranet caching space, whether you're in the U.S. or Europe, there's again two different value propositions. Internationally the transmission costs, backbone costs, are all very high, so the sales pitch there is installing cache and cap your growth in transmission cost and get a payback typically in three or four months. If it's in the U.S. where bandwidth is cheap, it's typically to help improve the user response time as a class of service differentiator.

Dale: Are international transmission costs one reason why the NetCache product did so well in Europe this quarter?

Warmenhoven: Absolutely. Yes, not just in Europe, but in the Far East as well. All of the international markets have been very, very strong in NetCache. And in fact, the U.S. market -- typically the U.S. is 60-70% of our total revenues, and in the NetCache arena it's 30%. The NetCache business has really been driven by this cost saving [in] international [markets].

Dale: Is NetCache -- is there a particular customer that's taking that up and is that a new product?

Warmenhoven: It's been available in the marketplace for about a year. We've introduced our second generation cache proxy server appliances in, I think around October of last year, so now about three or four months ago. And that's been, if you will, the volume producer in this past quarter, the new one. In fact we shipped over 100 units, out of 800 total for the company, of NetCache appliances.

Dale: Is there a certain sort of company that it is particularly useful for -- say you're talking about ECAD where you have a bunch of engineers accessing a data.

Warmenhoven: Well, on the NetCache side, the primary buyers are ISPs [Internet Service Providers] and multinationals running large intranets. The ISPs are the ones that pay the transmission costs -- so those have been the early adopters. I believe over time you'll see more and more of the large multinationals adopting caching. I'll give you an example -- the filer product really has been adopted very widely in the technical applications technology companies. But they're also starting to move towards caching. Cadence Design Systems is a good example. Cadence has been a large customer of ours for filers and just recently in this past quarter bought a cache proxy server to use internally and tie together their engineering teams. I think Ericsson at this point has about 30 of their sites that are intranet, and Ericsson has been another large filer customer over the years. They have about 30 of their development sites tied together with NetCache.

Dale: They have teams all over the world that share and update files every day, and I would think if you made one or two transmissions it's a heck of a lot better than accessing storage on the other side of the ocean a number of times a day.

Warmenhoven: That's right. And it's true that inside companies you generally get a better hit rate in terms of density; meaning, if somebody goes to get a design spec out there, somebody else working on the same project is going to go get the same spec, and so there's a higher probability because of the similarity of the information you're going for. I think it's pretty good payback on a intranet in a hurry.

Dale: For an ISP, does it save on peering expenses?

Warmenhoven: It does reduce the peering cost because it reduces traffic they try to access over the backbone.

Dale: And how else would it benefit an ISP? By increasing their available bandwidth?

Warmenhoven: Yes, essentially it caps the total transmission cost. These things will end up being in three or four different kinds of environment. If you put it in a POP [Point of Presence], for instance, where your customers are dialing into, you can reduce the long-haul bandwidth back to wherever your data center is. If you put it in your data center, you can reduce the amount of traffic you're putting on the backbone -- the Internet backbone, meaning the peerage backbone. And if you're in the hosting business you can use this as a way of capping the growth or the demand for servers behind the cache. The Tripod unit of Lycos uses our proxy server and they have probably six at this point, and their view is� their performance estimates [indicated] that one NetCache appliance can essentially avoid the addition of up to half a dozen Sun servers. Just by having one front-ending all the information, stuff that gets accessed on a repetitive bases comes off the cache and doesn't have to go to their hosting server.

Dale: Is there a certain size company that your products work better for? Say if I'm a small company, is the price-performance advantage not as large as I might see for a larger company?

Warmenhoven: In the caching space I'd say yes, there is a minimum. I don't know what that is, but in a file server space you probably have to have about 50 employees, roughly. The entry price for one of these systems is typically around $20,000 to $25,000 on the low end. The filers come in sizes small, medium, and large. At the low end it's probably $20,000 or $25,000. You probably need about 50 employees to justify it.

Dale: What's the advantage on the LAN [Local Area Network] in buying one of your products over buying just an NT box?

Warmenhoven: Well, I think it's faster, simpler, and more reliable, and generally more cost effective. Let's start with faster. In general, you can get performance out of these systems that's faster than a local disk. You probably have a file server -- an NT system or Windows system

Dale: Yeah, an NT system with H drives.

Warmenhoven: Ok, an H drive is your network drive?

Dale: Yes.

Warmenhoven: You probably don't use that all that frequently unless you have to.

Dale: I try to avoid it.

Warmenhoven: Right. So if I could give you the system that was actually faster than your local disk, you'd probably think twice about it because it automatically gets backed up and all the rest of that stuff, right?

Dale: Sure, I'd like to see that sped up.

Warmenhoven: And another problem that you probably have on your H drive is availability -- you can't count on it being there. So if I could solve those two problems from a end-user viewpoint -- and I can be very competitive on price -- and simplify the life of your system administrator, it really just makes it so he doesn't have to do anything. Not only would you be happy, but he'd probably be happy. And that's basically the heart of the sales pitch. It's been a real challenge getting the message out. Now I've got to tell you when I make claims like "faster than local disk," I wouldn't be surprised if you were a skeptic because I haven't met anybody who believes that when you first make the claim.

Dale: Well, I would take that as a verifiable fact so, you know, I'm not skeptical.

Warmenhoven: Most of the techies are, and in fact it is very verifiable, and a major component of our sales activity typically is giving the customer an evaluation system so they can prove it to themselves.

Dale: Is there a kind of a center of the market that you want to be shooting for in coming quarters and coming years?

Warmenhoven: I think of it as four separate markets, and I want a significant position in each. The Unix file server market has dominated [with] technical applications. And that's been our core strength. You know, the ECAD, software development, MCAD, and ISP set was kind of the core strength over the years. The second is the Windows market and the Windows file server market is one that we really started to focus on about two years ago. You know, we've build up, with Microsoft's help, basically a complete NT equivalent server. And that market is one that is the basis for the OEM relationship with Dell and Fujitsu. Part of our strategy to penetrate that market is to have Dell and Fujitsu resell this product for us.

The third segment is the database market. Storing Oracle data. or Sybase data, or Sequel Server data on our systems. And that's a fairly new one to us, but it's also a very large one, so I'd like to see a lot of growth go there. And the fourth one is this NetCache proxy server market, which is a brand new market and a fairly new product. The total worldwide caching market last year, my guess is, probably around $100 million. So it's a small and new market, but it's forecast to be several billion dollars in the next three or four years.

Dale: And the first three parts of the market are in the $10 billion dollar range that I've seen?

Warmenhoven: $9 billion.

Dale: And what's the growth rate one could assume for that market?

Warmenhoven: Well, let me give it to you two ways. The over-all nine billion, I think is growing about 20% per year.

Dale: In units or dollars?

Warmenhoven: In dollars. There's a segment of that market that you've got to think of as being traditionally addressed by general-purpose computers. Your sever down the hall is probably an Intel architecture running NT.

Dale: Yes.

Warmenhoven: So the system down the hall is in that $9 billion. If you look at what portion of the $9 billion is being addressed through appliances -- that is, machines designed just for that task -- it's about 10%, or $900 million. So our share of the storage server appliances market would be roughly about 33%.

Next: Part 2 of the Interview