Hewlett-Packard appears to have developed an extremely viable strategy by forming alliances that co-opted former rivals such as Microsoft and Intel into long-term partnerships. But that strategy requires at least a three-year commitment before the company will see real material benefits. In an interview with IDG's Sandy Reed, Michael Vizard, and David Pendery, HP Chairman Lew Platt talks about what factors need to come together to allow HP to move ahead as the venerable systems vendor approaches the new millenniumIDG: In the wake of the Microsoft alliance, many people feel that Windows NT has not completely lived up to expectations in the enterprise. There is some concern that HP hurt itself by over-committing to NT at the expense of HP-UX. What is HP's current strategy?
Platt: Our strategy is very simple, but people try to make it incredibly complicated. I think we are following through on being the world leader in Unix systems, which is still the choice for large, complicated, highly reliable enterprise systems.
But that doesn't mean that we don't intend to be strong in the NT space and in the Wintel space as well. We're a big enough company to do both.
Some of your rivals, such as Sun, would argue that you've simply sold out, there by hurting your Unix business. How is HP's Unix business these days?
First of all, Sun's got a problem because I think they've got the wrong strategy. It's a simple strategy, but it's not the one that most customers really want because customers want to bring these two environments together. As for abandoning Unix, nothing could be further from the truth. Unix sales are healthy and growing, and we're increasing our investment there. Sun just somehow refuses to believe that a $47 billion company can do two things at once.
We've taken a lot of heat about abandoning Unix, so we've tended to tip the marketing messages back toward Unix just to let the world know that we're still alive and that we care about it.
What is HP's relationship with Intel like these days, given the cooperative Merced effort to build a new high-end processor architecture to rival mainframes?
The relationship with Intel is as good as it's ever been. A lot of people said HP must somehow be out of the loop because we appeared to them surprised by the announcement of Merced delays. We weren't surprised at all.
We know exactly what's going on, but by agreement we gave Intel fundamental control over the marketing. We're moving along on the technology, but it's probably not all that unexpected that we had some delays.
What implications does Merced have for HP's existing PA-RISC architecture?
We have had a plan from the beginning that includes other generations of PA-RISC processors. We said from the very start that we need to have a smooth overlapping in order to have a smooth transition from PA-RISC to IA-64. We still have several generations of PA-RISC planned. We may or may not do them all. It simply depends on how quickly we're able to get the IA-64 program done and basically how much of that overlap we need.
Has HP been tracking the rise of Linux?
It's got a lot going for it and it's coming up really fast. The problem it has to solve is that you need to get some real-world applications supported on it as a primary core, not some guy in a back room.
At the point that those guys embrace it, we'll have a Linux strategy. Big-time ISV support is really the key issue.
With HP's huge base in manufacturing systems and the enterprise in general, the rise of Internet applications must be having a profound effect. What does HP need to do next to capitalise on the Internet?
We have developed a lot of core e-commerce technology. What we haven't done is rolled out the marketing message in really coherent and easy-to-understand terms. In all honesty, today we have a lot of very good tools and parts and pieces. What we need to do next is package this up.
We haven't really done a particularly good job of rolling out complete solutions, but you'll see that shortly from HP.
A lot of people believe that the Internet is actually fostering a return to centralised computing because it has become easier to distribute data access rather than the actual systems. What's HP seeing with this trend?
We are definitely seeing a return to some centralisation among many of our clients. But that doesn't necessarily mean a return to the mainframe. It may be a return to more control over the high-end servers, and even some control over the clients in the client/server environment. We don't think the alternative is to go back to a mainframe because then you have DB2 and CICS.
Well, you don't get Oracle support for a DB2 CICS mainframe environment for obvious reasons.
Speaking of PCs, what is HP's current thinking about direct vs indirect sales?
We don't have a lot of religion around direct vs indirect. Obviously, we've built a lot of our position and strength using the channel.
You cannot, on the other hand, ignore in any way what Dell has done. Their model has cost advantage, it probably has some inventory advantages, and customers like it.
But rather than try and go duplicate the Dell model, we are going to go direct with a class of customers over the Web. If we're going to be successful serving those customers, we're going to have to do it directly.
In the PC space, the one that you have to keep watching and learning from is Dell.
We're pretty happy with our performance against virtually all the other PC competitors, with the exception of Dell.