sn -
I doubt that CPQ will purge Alpha but I would expect Alpha to be correctly segmented, something that has not happened yet. Alpha is a superior processor for high end systems, the best in the business at the moment. Sun has made a business with a proprietary chip, the SPARC, which is a generation behind Alpha and has half the performance, and they do it on a small number of units. CPQ IMHO needs to develop a focused high end position for Alpha (the Himalaya line, large SMP Alpha systems) and NOT position Alpha in the volume segment. For the next few years, that will probably extend down to 4-way servers since there will not be a viable 64-bit Intel offering until McKinley, probably in mid-2002, and even at that time Alpha will have about twice the performance.
At the same time, CPQ needs to develop leadership in the 64 bit space so that they are the partner of choice for Intel's IA64 offerings. Even though Alpha will have better performance, Intel will certainly be the volume leader with IA64 and CPQ needs to own that space.
Segmentation of the product lines is important so that CPQ does not confuse customers, waste time on redundant engineering, and present partners with a conflicting roadmap. This would also allow CPQ to control the costs of Alpha development by spending on the areas that lead to sales. Sun is doing it, CPQ can do it too.
K7 is interesting but Intel is hard to beat. Intel will continue to dominate as long as they can leverage their ability to spend more on fabs, which in turn gives them better process capability, which gives them more advanced products in bigger volume. They just have to keep the treadmill spinning. Intel's problem would be if the market somehow became content with current processor technology, which seems unlikely to me any time soon. |
