Dave Wilde: Coming from the R&D side, one of the issues we have right now is that we dovetail with the rest of HP. The lowest-end boxes right now on the Unix side are PCI-based systems. Right now our [3000] IO system is not designed in such a way that it could run on PCI. One of the biggest projects we have in the lab right now is to move to a set of next-generation boxes which across the board have PCI technology inside them. That means a major rewrite to our IO subsystem.

Over time there will be a lot of benefit [to the installed base] in that area. But the real benefit will be that it gets us in line with where the mainstream of HP is. That will allow us to scale much higher on the high end and reach much lower on the low end in terms of cost-competitive boxes. That’s an evolution toward what we’ll be able to do on the IA-64 boxes.

The timeframe for the first of the PCI-based systems is late 2000 to early 2001. We are working very aggressively to get those systems out as soon as possible. Our first focus is going to be on a very high performance system. Then we’ll be looking at scaling up and down from there.

Vicki Symonds: I think we can take a look at the pricing of our low end and see what we can do. Obviously we have some constraints there in terms of how low we could go.