>With I2O, could it level the playing field such that CPQ, DELL or whoever
>could start barging into SUNW's domain in a big way?
Dave's response to this question is absolutely correct, but it you were to reduce it to one word it would be "yes".
It has been a consistent, but unhappy, theme on this thread that NT is an enormous threat to SUNW, and that outcome continues to accelerate, despite lack of maturity of Windows NT. I2O does help level the playing field and as such simply speeds up the process.
Swashbuckling SUNW became a favorite of the engineering crowd with, at the time, awesome workstations that could be expanded seemingly forever and that had a natural ability to network, inherited from the famed Stanford University Network (SUN). In the nineties, the line expanded into robust and powerful servers.
Meanwhile NT, designed as a new, proprietary operating system by DEC, came into being by mopping up the low-end. First to go are the Unix workstations, which is happening at breakneck speed today. Second to go are low-end, non-mission critical servers, which also is happening as we speak. The high-end servers will go last, but very slowly, as Unix vendors try to keep ahead of MSFT by being first with 64-bits, a greater number of multi-processors, and traditionally, much better I/O. The latter feature is under direct attack from the PCI bus and I2O. As these become required standards for I/O, and as Intel/HP introduce the 64-bit processor, only serious multi-processing, operating system robustness and application programs separate high-end Unix servers from NT.
In case you haven't noticed, Intel is the world record-holder for multi-processing. It recently won this distinguished award based on its Cougar Operating System, which is the commercial name for what Sandia Labs internally calls "Puma". By no means does this suggest that Intel necessarily is going to compete against MSFT in the OS arena. But it does mean that the ability to efficiently incorporate multi-processors in a server is not an inherent weakness of the WinTel crowd.
Also, I wouldn't put much faith in lack of available application software. By now most important software for the NT probably has already been ported or written, and software developers will not hesitate to complete the job - because that is where the market is growing. For example, WIND makes no bones about the growing importance of Windows NT/95 for Tornado versus Unix.
In the final analysis, the only thing standing between NT and total server domination is robustness of the OS, and in time even MSFT should get that right.
The first computer companies to take extreme hits of course were the so-called proprietary computer firms like Data General, Prime, DEC, Unisys, NCR and even IBM. Besides WinTel, full-fledged survivors of the first round were Unix companies like SUNW and SGI, and the Unix components of IBM and HP, for example. Now we see that SGI is showing business weakness and is under repair. SGI has always been exceptional because of its graphics capabilities, extended by acquisitions of Alias and Wavefront in direct response to MSFT's acquisition of Softimage, which caused the numbers in their respective business to collapse. Today there is nothing that differentiates an SGI graphics workstation from an NT - even to the point that SGI intends to provide NT workstations starting probably next year.
SUNW will continue to enjoy considerable success in the server business, but not without the constant pressure of WinTel, and not by any accident. Over a year ago Intel pointed out to analysts that the workstation and server business were the only opportunities to grow market share. As a result, for some time I have viewed SUNW as a risky investment, and continue to do so despite its recent doubling from the time this thread began and SUNW was first discussed. (It has also been pointed out that stock prices almost always increase for long periods after it is first observed that a business is facing serious problems.)
Java is not enough to save SUNW, and especially not against an intelligent foe like MSFT. Even if Java were wildly successful, which could happen, I don't see how SUNW benefits sufficiently to offset the inevitable slide in the Unix business.
SUNW is a great company that unfortunately found itself in the computer business in the old-fashioned way, vertical integration of hardware and software. That business is obsolete. I think you can sense that Dave is sad about this, and so am I. But life goes on, and it goes on with the next wave of computing, embedded systems.
Allen |
