Adaptec: Friend or foe?

Robert Stephens' mission is a variation on the old saw, "the whole is greater than the sum of its parts."

As chief executive of Adaptec, Stephens has supervised the company's shift from a maker of components used for data storage purposes to a builder of complete storage systems.

Under Stephens' watch, Adaptec has gobbled up a couple of companies that make networked storage devices--which hold data separate from computers to allow for more efficient use of resources. Adaptec bought Eurologic in 2003 and snapped up Snap Appliance in July.

At first glance, the move to make whole systems is a curious one: Why would Adaptec want to threaten the supplier relationship it enjoys with companies like Hewlett-Packard by competing against them? And why, in the Snap deal, enter the so-called network-attached storage market when heavyweight Microsoft is making a major push there?

New revenue opportunities, of course, in the growing data storage market. Adaptec earlier this month predicted that it would bring in more than $200 million in storage systems revenue for the year ending March 2006. That's a hefty chunk for a company that overall took in $452.9 million for the year ended March 31.

And Stephens says his company can navigate the challenges. As he sees it, Adaptec can continue to team up with the big guns of storage by building more-complete products for them. Indeed, the company recently announced that it is supplying technology used in new storage products from IBM.

In a recent interview with CNET News.com, Stephens discussed Adaptec's goal to be both a parts and systems company, and hinted at a possible partnership with Microsoft.

Tell me about your strategy in moving Adaptec from being a components company to a full-fledged storage system seller. It seems kind of counterintuitive. You're going up against big guns like EMC, IBM and HP--even Microsoft, with its network-attached storage effort. What's more, you risk upsetting partners in your components business. Why is this the right approach for Adaptec?
You have to go back in history a little bit. When most people think about Adaptec, they think about it as a hardware-centric company, and that has historically been true. But even going back five years, at least, you find that the primary engineering work inside of Adaptec is software. Two-thirds of our engineering effort is really around software.

It's a very natural extension to move from a SCSI host bus adapter business to starting to manage storage within a disk array. (SCSI host bus adapters allow servers to use the small computer system interface for exchanging data.) So more than seven years ago, we staked out a position in RAID, which is all about software, really. (RAID refers to redundant array of independent disks, a technology to protect data against disk failure.)

So then the next natural extension to that thinking was that we were very host-centric, in our view--that is, managing storage that was either in the server or directly attached to the server. So it made sense to manage spindles that were part of either a storage fabric SAN (storage area network) or some kind of external environment away from the server.

What makes this unique is that we are the only ones that come out of external storage with a host view. You mentioned EMC as an example. EMC really doesn't have a host footprint. They don't sell to Dell or IBM or HP.

What about some others? HP is certainly a competitor, as is Dell. They both are server vendors.
Right. And so the first acquisition that we did outside of the RAID area to facilitate our move into external storage was Eurologic. And when we did the Eurologic acquisition, I talked with a number of our OEM (original equipment manufacturer) customers, because I wanted to make sure, to your point, that they were comfortable with our direction.

They were very comfortable--both IBM and Dell, to be very specific. HP was a little equivocal, to be candid with you. Six months later, when HP looks at it, it says, 'Well, this makes a lot of sense.' What it was looking at wasn't immediately obvious to it, in the context of the architecture. What we're really taking about here is not proprietary external storage, and that's what's different.

To give you an example of this, if you go onto the Intel Web site, there is a white paper about Adaptec's approach to external storage, using Intel's (input-output processors) and standard Intel architecture. We were the first ones to do that. We can save our OEMs 30 percent to 35 percent just in development efficiency--in being able to move products seamlessly from the host environment to the external environment.

So you are claiming that you guys are good at using a very standards-based approach of putting off-the-shelf components together in an inexpensive way with some server expertise behind it.
Absolutely. And what's magical about it is that the same RAID code and the same storage manager run from the component level all the way through the external platforms. So if you are an IT manager, you bring up your screen, you can look at your storage array--both direct-attached storage and fabric-attached storage--and manage it in precisely the same way.

You guys are talking a lot about going into the small and medium-size business market, but what makes you think you know what it takes to sell products there?
If you take a look at the last couple of years, maybe go back four years, there has been some wreckage in that space. A number of companies have looked at the market in the context of the (distribution) channel, and the thought process was maybe as pedestrian as, "We will build the product, and we will provide it to the channel, and they will come." And no one came.

We did not acquire Eurologic for a distribution channel footprint. We acquired it for an OEM footprint, because 95 percent of its business was really with OEMs. Now, you might think because we have a very significant channel position and a terrific brand that it would be a natural corollary for us to drive storage array-based products in the channel. But it's a different set of VARs (value-added resellers) that buy this product than our traditional VARs.

That's why we acquired Snap. We felt that its footprint in the channel would be a tremendous advantage. It has about 1,200 to 1,500 VARs that buy system-level products. It has a terrific brand. It has been No. 1 in NAS (network attached storage) for four years, in terms of volume. Snap was exclusively a distribution channel business. It did not even look at the OEM side. We think that that is really the key to major success--having a presence both on the OEM side and the channel side.

When you are saying Snap was No. 1 in NAS volume, can you give me a little more specific data?
It has well more than 100,000 customers. It shipped 150,000 units over, really, the last four years.

So then, in other words, it is not selling higher-price NAS devices but lot of smaller, cheaper NAS devices than, say, Network Appliance or EMC or HP?
The simple answer is yes--that's where it started. Then it started to move up. It's a lot easier to cannibalize the competition from underneath than to scale your architecture down. Scaling architectures into a much less-expensive bill of material and all that is a major undertaking.

So with this Snap product, you think you can eventually make enhancements over time to compete against an EMC Celerra or the Network Appliance machines?
Not exactly. Snap has today a combination product--the (block-level and file-level data) product that Network Appliance has. Frankly, what we would prefer to do is work with Network Appliance, not compete with it.

Why would NetApp want to work Adaptec?
There are configurations, certainly, and elements of architecture that would provide an advantage to them. I don't want to single out NetApp, but we see across the broad spectrum opportunities for "private label" products, whether it's Eurologic-based or Snap-based. We have been working aggressively on that front. That takes some of the risk out of the equation, from my standpoint. I would much sooner use a tier 1 sales force than to have to do all of this work organically. And so I am going to continue to walk on both sides of the street.

And now we get to back to your fundamental point on components. Why does this make sense for Adaptec, and why does it make sense to the customers? It really has been about being able to provide an end-to-end storage solution to the marketplace. And when I talk about marketplace, I am including OEMs.

Doesn't the Snap move also put you up against some potentially strong competition from Microsoft and its attempt to get a bigger storage presence in the network-attached storage arena? How are you going to fight Microsoft there?
I have no intention of fighting Microsoft. We have always been a strong ally with Microsoft.

But Snap has a proprietary operating system that, in a sense, competes against Microsoft and its Window Storage Server 2003.
Sure, it does. Snap's operating system is Linux-based--it's called Guardian OS. But there is nothing that says we can't offer a Microsoft-based solution as well.

Can you say anything about that right now?
I can't say anything about it. But I don't like to stake out a proprietary position. Walking both sides of the street here, for the Linux users and for the Microsoft users, makes a lot of sense to us. We don't see a reason that we should take sides.