Part 1, Barrett unplugged: 'Put that infrastructure in place'

"110"="" hspace="3" vspace="1">Intel CEO Craig Barrett oversees a nearly $30 billion empire that leads the world in the production of microprocessors. Recently, Barrett spent some time with ZDNet Editor in Chief Dan Farber and Tech Update Editorial Director David Berlind at the company's headquarters in Santa Clara, Calif. In the initial installment of this four-part interview, Barrett offers an IT prescription, describes how price/performance is a key metric, and tells how innovation shouldn't fall victim to a sluggish economy.

Tech Update: According to some estimates, by the end of this decade IT expenditures are going to account for about 50 percent of an enterprise's operating expenses. Given that, a gross miscalculation of a long-term IT strategy could mean the difference between life and death for an enterprise. So, if you woke up tomorrow as the CIO at one of the companies in our audience, what would your IT road map look like?

Barrett: Well, I think everyone is pretty much in the same boat. Everyone is looking at the infrastructure that they have. Most of us have infrastructures that have grown up over time. You have separate databases, separate applications that'll talk to one another. If I was to wake up tomorrow [as the CIO of another company], I'd say the most important thing is to put a kind of a uniform, horizontal infrastructure in place--which is seamless hardware, seamless databases, seamless business processes. And, then drive my applications off of that kind of pancaked, horizontal baseline. You look at the issues that we're facing today as we make that conversion, and one of the major issues is how to get databases to talk to one another. The most graphic example in the world is the United States government and their whole response to terrorism in the 9/11 event. You know, if everybody has their own database, and nobody talks to one another, and there's no coordination … that's not much different than a lot of corporations today. So, if I was an IT guy, I'd make sure that I'd put that infrastructure in place. I'd pick out the most critical applications to make me competitive in my particular business model. Obviously, that depends on the channels and products, and the geographies in which I would be operating and a whole bunch of things. I'd pick the most critical business applications associated with the Internet that are going to make me competitive and put my wood behind those couple of arrows--the ones that are the most critical.

Tech Update: What do you mean by "put your wood behind those arrows?"

Barrett: My investment.

Tech Update: So, what kind of investment would you make, aside from making sure they talk to each other?

Barrett: Well, that's an open-ended question. I'd put enough investment to make them successful. If the most important thing I had to do was to have a manufacturing, planning, and response system to respond to market demand, then I'd put my wood behind that arrowhead and make sure that I got that in place. If my most important thing is customer relationship management, then I'd put my wood there. If the most important thing I could do is to take information throughout the corporation and combine it into a common-designed database to create my next generation of products, I'd put my wood behind that. So, until you lay out the whole problem, I can't tell you how much wood to put behind each one of those things. I need the gory details first.

Tech Update: Provided a lot of corporations have this transition to go through, like the merging of databases, why should Intel be a partner in that process?

Barrett: First of all, we're probably a poster child for other companies. We're a manufacturing- and engineering-oriented company. We're also a marketing-oriented company. So we have our fingers in a bunch of different pies. We have been aggressive in this space in terms of our own learning and our own investments so far. If you look at us in terms of using the technology to sell product, to buy goods and services, or to educate the system integrators around the world to pick our product and move it forward into the marketplace, we've been very aggressive in those areas. So, from an experiential standpoint, we're experienced in what we are talking about. If the question is, "Why should Intel as an infrastructure supplier, or a capability supplier, be brought to the market?" then there's only one reason that we should be brought to the market: if we bring some benefit of value to people. We're the price/performance leader in just about every situation, especially when you start thinking about pancaking into common databases, common hardware infrastructures, common business process capabilities. Price/performance is the model that we live, eat, sleep, and breathe--and have for 20 years.

Tech Update: Is price/performance the top metric? What about other criteria such as reliability?

Barrett: That's part of a total cost of ownership. If you go back to the mid-1980s when quality was the major issue in the semiconductor industry, the Japanese brought forth quality as a competitive advantage. The whole world said, "OK, that's interesting." Everybody has to have quality to compete. It is not a competitive advantage per se--it's a necessity to compete. At the enterprise level, reliability is pretty much a necessity to compete. You don't put databases in place, or data centers in place, and say, "OK, 95 percent up time is good enough." We get people who are arguing whether you are five-nines, four-nines, or three-nines in terms of reliability. It's kind of a rounding error because everybody defines their own capability a little bit differently. They're probably all four-sevens. So, that's a basic requirement that we have in the system, or you don't play.

Assuming you pass that activation barrier to get in the system, then the next question is, what's the total cost of ownership? What's the price/performance? That's where you see the world going today. The world is not rich. The IT dot-com bubble has burst. The Internet euphoria is gone. But, since everybody recognizes the Internet is a critical part, everybody has to invest to take advantage of it. But, before they do that, they must look at return on invested capital. That's why we are looking at total cost of ownership. Price/performance, therefore, gets to be a very, very critical aspect of what people are doing.

Tech Update: At the Intel Developer Forum you said that to get out of the recession we're in, it was necessary to continue to innovate. You said that the engine of innovation would be the Internet and the convergence of computing and communications. In the context of TCO, how do we get from where we are now to that arena of innovation?

Barrett: I'd have to parse that out into the different levels. It's different if you're a semiconductor manufacturer vs. a system manufacturer vs. an end user. In the semiconductor space, the average lifetime of a product is a year or so. Unless you innovate, you don't have products to sell into the marketplace. If you go up to the next level of system manufacturing, or the next level, which is the end-user capability, then the cycle isn't going quite so fast and that ultimately affects competitiveness. The question isn't whether a system has a 2GHz or a 3GHz processor. The question is why are systems being scaled back to match end-user capability? Innovation has to occur at that higher level, too (from the system end-user standpoint) for the very simple reason that it is your competitive position in the marketplace.

Here's a very fundamental baseline case. Go to any country in the world and talk to the business and government leaders, and what do they tell you? My IT infrastructure determines my competitiveness going forward. Do you want last year's competitiveness or next year's competitiveness?

Tech Update: But reinvesting takes money and we've seen a worldwide slowdown in technology acquisition.

Barrett: You've seen a slowdown to some degree in mostly enterprise investment.

Tech Update: So, what's it going to take for companies to see the connection between IT investment and competitiveness?

Barrett: Profitability.

Tech Update: Profitability by the end-user companies so that they can buy more products?

Barrett: Everyone gets into a difficult time, and they get the bunker mentality, and they tend to put budgets in place which are consistent with what's going on in the outside world. They cut things that are discretionary and upgrading ThinkPads is discretionary. Or, maybe in your mind, upgrading something else is discretionary. When you start to drive is when you start to recognize, "Uh-oh, I put that off and now I'm at a competitive disadvantage because my sales guy or my reporters don't have the competitiveness to compete with the other guy because they've got to plug their systems in [and the competitor is still running on battery power]." Or maybe you'll suddenly realize that your guy doesn't have a wireless connection while the competitor does and he can communicate any time, any place, and to anyone. Meanwhile, your guy is looking for a phone jack. Then, you're going to make that investment. Or, maybe you wait another year because you really don't see the sense in it and the next thing you know, you're really behind.

If you want the best example of what happens in terms of long-term competitive disadvantage, look at Japan. The lack of investment in Japan's IT infrastructure and the lack of modernization of their capability has been one of the major contributors to the inefficiency and stagnation of that country over the last decade. That may be one of the reasons that, if you go to Japan today, you'll see the NTT DoCoMos and the other communications companies installing [fiber-optic networks] to the home everywhere. It's a long-term competitiveness issue.

Tech Update: So, the price/performance message seems to be getting some traction. Intel has done a good job of selling the top big iron providers on IA-64. A lot of companies like IBM, HP, and Compaq are making the transition. Are you counting on them to get you into the data centers, where the enterprise IT people will recognize that Intel's price/performance--coupled with battle-tested operating systems--will contribute to long-term competitiveness?

Barrett: You have to look at the arrangements and the agreements that we have with companies like Compaq or HP, which separately committed to the Itanium processor family. Both committed to make the conversion. What happens when they combine, if they combine? If it goes forward, then there would be some consolidation in the new Hewlett-Packard product offerings. But, jointly, they are both committed to [IA-64] as the hardware platform of the future. You've got Microsoft, which is committed to this with 64-bit Windows, and you've got support from the Linux community as well. So, it could be pretty much what the end customer wants, you know? Will the financial community go with Windows? Will they go with HP-UX? Will they go with Linux? Well, if you go wander around Wall Street, you'll get different answers from different people.

But, I think you'll see some combination of all of those systems out there and that's what will drive [IA-64] into the data center. By the way, it's not just the support of the OS. It will be the price/performance of what you put into the complete hardware/software package. Another thing is that it is also available from multiple suppliers, so there's competition out there. You are not locked into one manufacturer. The bottom line though is that if it runs 50 percent faster, and it costs 20 percent less than what the other guy has, then that kind of gives you a 10X advantage. Why not?