Part 3, Barrett unplugged: From the pocket to the data center
Summary
Topics
In Part 3 of his interview with Tech Update, Intel CEO Craig Barrett explains how Intel plans to leverage XScale to squeeze more power onto more small devices, why the "price/performance advantage" is key to getting more major customers to convert their big iron solutions to Intel architecture, and why he's not afraid of being "hammered" by AMD.
Tech Update: The handheld market, which includes a lot of consumers, is certainly big business. People are also moving to smaller laptop configurations. Intel is becoming a major player in those spaces. How do you see that part of the business evolving in terms of percentage of revenues, as well as the ability to get more processing power into those devices?
Barrett: Well, that is our strategy in a nutshell. When we coined the term "personal Internet acquired architecture," we set it up with some very simple fundamentals. Handheld devices have two functions: one is communications and the other is computing. You want those two things to continue to grow and integrate together, but you don't want either capability to be hampered by the other. We've always said that the solution is to put in a general-purpose computing engine and keep it somewhat separate from the communications engine so that you could upgrade the computing engine without having to go back to the FCC and then having to upgrade the whole thing.
So, Ron Smith and his group at Intel is following that exact strategy with our second generation of StrongArm, which we named XScale. The idea is to combine that with our flash memory and to bring the best in compute-memory performance into that space. We also have some fledgling baseband capability in there for the communication part of it. Eventually, we'll try to integrate all that. But currently, it's basically shoving the price/performance leader from the computing standpoint into that space.
If you look at the Pocket PCs, StrongArm is almost a clean sweep. I presume that everyone will upgrade to the XScale products we're introducing. So I think we have a good leadership position or, at the very least, a good strategy position. The real issue now is taking people from their old approach, which was to do ASICs where they combined compute and communications stuff into proprietary chips, and start to use standard building blocks. This speeds up time to market and lowers cost. It also gives you better performance. We've been on this tack for a year or so and it represents a huge growth opportunity for us.
Tech Update: How will you make the merits of Intel's newer products better known, in the same kind of way that Microsoft is out there with its agility campaign trying to kind of convince people that Microsoft is more relevant for the enterprise now?
Barrett: As you probably know, we do spend a few bucks each year on marketing. The "Intel Inside" brand is reasonably well-known around the world.
Tech Update: How about for the other segments where that hasn't gotten as much traction? Perhaps the mini-mainframe segment that we were talking about before, where you need more of a big iron approach?
Barrett: Well, the big iron approach is going to be sold more on a cost-of-ownership and price/performance capability. I'm very, very comfortable with what we bring to that part. When I go to Wall Street and visit with major customers, and see the transitions taking place and the design wins going down, I'm very comfortable with the capability that we bring into that space. We don't need so much of a branding campaign or an advertising campaign; we need more and more major customers converting their big iron solutions to Intel architecture.
We just finished our analysts meeting where we reviewed last year's goal to get 10 marquee company design wins for the Intel architecture. In the big iron category, we ended up with 167. Now, we're trying to double that success this year and expand it going forward. So, step-by-step, we'll put the architecture out there, and get the OS support for it, and get the applications on it. And then we'll get the end customer to look at the price/performance advantage.
The people who put out the best technology at the best price point win.
Tech Update: Speaking of price point, AMD has made a lot of noise about Hammer. Does it represent any threat to your business?
Barrett: Well, I don't think it's any threat to the enterprise business. I mean Hammer is nothing more than an address extension off of a standard 32-bit machine. It doesn't have data center class capability. That's what the whole Itanium processor family was built to be. So, I don't look at it as a threat there. I don't know where [AMD CEO] Jerry [Sanders] is positioning the product these days. I mean, he just gave it a new name, Opteron. If it has a home, it has a home in those applications that are not data center applications but where you may need that increased addressability in the short term. But that's a very limited number of applications. Microsoft has announced support for it. But, as far as I understand it, the real support for it comes in Microsoft's next generation OS--the one code-named Longhorn--which is a couple of years out. I think there are some interesting applications on the desktop for that sort of increased addressability.
Talkback - Tell Us What You Think
The best of ZDNet, delivered
ZDNet Newsletters
Get the best of ZDNet delivered straight to your inbox
