This is the first realistic analysis of the Intel/Sun deal that I
found:
news.com
I think Sun's customers will be more confused about the deal
than anything. As the author points out, Intel Solaris has
been available since 1993 and has gone basically nowhere.
What I don't understand is how Sun can time this now. It seems
that UltraSparc III will arrive at the exact same time. How to
market or differentiate? Good that it is their problem.
But not all are as downbeat as that last article. Check out
the 12/16/97 Wall Street. Our friend Dean Takahashi ( a Sun mole
working at Wall Street, see my landmark piece following this post
to confirm such a statement ) writes a much more glowing article:
"From Sun's perspective, the deal also helps advance a dual strategy
for its future. On the other hand, the company will still make and
sell computers based on its own microprocessors, known as UltraSparc,
and which will run its Solaris operating system. On the other hand,
Sun, of Mountain View, Calif., will also offera a version of Solaris
to run on Intel-based microprocessors, vying for the same customers
who might opt to use a future version of [ NT ]."
Look how wonderful! But how was Digital painted recently when
they announced their agreement? How about a nice snippet:
news.com
Apparently, under the proposed terms of
settlement, Intel will pay Digital consideration worth
$1.6 billion, with nearly half of the total coming in
the form of soft dollars, processor discounts, and
development and manufacturing assistance, a huge
amount of nonmonetary consideration that would
make sense only if Digital were planning to move
away from its own Alpha architecture and toward
Intel's 64-bit Merced chip.
"It makes sense only if Digital was planning to
convert to IA-64 [Intel's 64-bit architecture] in the
future," said Linley Gwennap, editor-in-chief of The
Microprocessor Report. "It doesn't make sense
that Intel would be 'fabbing' [manufacturing] the
main competitor to Merced for an indefinite time.
You see it is just peachy that Sun can pursue a "dual strategy" but
for Digital, "it just doesn't make sense." Marginalization again.
Sun is the darling, Digital is the wanna-be.
I bet we will see .25 21264 Alphas real soon (4 months or so). Here
is another Alpha partner that is ready to roll at .25 micron:
infoworld.com
When Samsung mentioned a 700 MHz 21264KP part they are getting ready
for (at .25) the press went gaga for a while. I bet what is happening
since the 21264 is so long delayed the .35 parts won't be around long
before the .25 parts show up. At .25 Tom we know that they should
come in much faster than Merced I and a year ahead. How much faster?
I don't know. I read a while a go that .25 should be 50/70 on int/fp.
Search www.byte.com for 21364 for that article.
Sun is on a me-too kick here. I bet the pinch from sliding workstation sales have got them more than a little frightened. They
certainly can't support NT (not with the animosity between Mr. Bill
Gates and Scottie McNeely). That leaves Sun little choice in the
Wintel duopoly but to support Intel. The world will be doing NT
everywhere except the Enterprise in 2-3 years and then in the Enterprise after that. Sun really has nowhere to go in about
4 years. Digital is smart in that they are focusing on services.
Somebody has to support NT. It certainly isn't Dell or Gateway.
Rob |
