That sounds like a different story. The one I read seemed to say their complaint was that Sun wasn't recruiting application software vendors fast enough.
Intel has been suspicious of all of Sun's efforts throughout the '90's and even before, beginning with the failed 386i or "RoadRunner" machine, an Intel-based 1988-89 Sun workstation with a proprietary (non-PC) architecture that flopped resoundingly, embarrassing Sun and infuriating Intel. Their relationship never recovered.
Since Solaris x86 started up in 91 or 92, Intel was always making noises about a variety of things, as reflected in the ZD puppet press: would Sun get enough applications, would they deliberately slow down the Intel version of Solaris just to make Intel look bad and SPARC look good, would they get enough unit volume to avoid another embarrassment, was Solaris x86 just a trojan horse to allow the Sun sales organization to get into an account with the promise of cheap hardware and then pull a bait-and-switch to sell them Sun hardware instead, yadda yadda yadda. In other words, was Sun doing the whole thing to make Solaris a Windows competitor, a major O/S contender on commodity hardware (as Sun claimed), or were they doing it just to muddy the water and confuse the issue in the high end of the Intel software market?
Problem is, Sun never really put forth a vigorous answer to those questions. Solaris x86, originally handled out of a separate software operating entity within Sun (remember SunSoft?) never took off, never got a lot of SPARC applications moved even though the move was trivial, never got the real backing of a large Intel hardware OEM (which was an insurmountably limiting factor) and was eventually collapsed back into the overall "big Sun" software organization when SunSoft ceased to exist at the end of 1997.
What was it all about? Sun did the port, it worked, it performed well (it obviously wasn't sabotaged as Intel pretended to fear), it's still used by thousands of customers and a lot of schools, since it's now free. The technical part wasn't really the issue. It was the rest of the stuff, the marketing, applications, OEM partnerships, market share growth, etc. that never materialized. It's hard to imagine that it ever made Sun a dime, though they never broke out numbers.
Could Sun have made it happen with a bigger investment, and did they really care that much? No one knows except McNealy and Zander, and maybe not even them.
Once again, in the case of Itanic, the porting part is the least of it. The idea of technical sabotage is 100% baloney. Sun's not likely to spend millions (and a port definitely costs millions) just to make one of their own products look bad and piss off everybody in the industry, despite Intel's paranoid wives' tales. In order for Solaris Itanic to make Intel look bad, it also has to make Sun look bad: there will be other Itanic operating systems (eventually, maybe).
But was there political infighting inside Sun that kept marketing resources directed elsewhere? Was more attention paid to higher-margin products especially after market share for Sun servers grew healthily between the time the Itanic port started and the time of the chip's release, which will end up involving over a year's slip by Intel (not by Sun)? At least one of Sun's original OEM's (Siemens) abandoned Merced in favor of SPARC because of Intel's slippage, not because of anything Sun did. Was Sun constantly battling hostility and suspicion from Intel?
Sun is a chiefly a hardware company, in that their revenue comes mostly from hardware. More and more, they're touting free software anyway. The part of Sun that builds computers is, and always has been, the dog and the software part is the tail. You can draw a lot of conclusions from that. A lot of things, especially marketing operating systems for other people's hardware in a hardware company, can get screwed up without a nefarious conspiracy. Just look at IBM, which has no peer in that department<g>.
I never understood either Solaris x86 or Solaris Itanic and still don't.
--QS |
