Why does Intel support Linux?

Chris Gehlker wrote:

> Thanks to you and everyone else who responded. You have given me lots of
> food for thought.

As you did for me. :) I enjoy looking at these strategic decisions and
figuring out why companies do what they do.

> But AFAICT it is exactly the Wintel association that is responsible for
> their success. It's not like x86 was the best chip. There have been much
> better chips around that just vanished because they couldn't run
> Windows. And that includes better chips from Intel. Remember the i960? I
> do take your point about Linux giving them entrée into the high end
> market that was dominated by Solaris, AIX, HP/UX and a bunch of
> proprietary *nix. But having a foot in the door and successfully
> competing are different things. I think that space was HP's for the
> taking and they just blew it

Sure, the x86 was considered inferior to lots of RISC based chips.
However, wouldn't you agree that effectively Microsoft is tied to chips
that are either ia32 or backwards compatible with ia32 (ie the AMD64
chips)? As long as Intel stays competitive with AMD, they will maintain
that market. If Intel wants to expand its revenues and profits, then
they truly need to expand big into other areas -- high end servers,
embedded systems, etc.. in many instances, Linux allows for this. Sure,
they will need to be successful in order to get market share, but they
at least need their foot in the door.

> The fact that Intel has leverage against MS and does not have leverage
> against Linux was precisely the point of my question. It does raise an
> interesting question though. How portable is Windows? Do we have any
> assurance that MS doesn't keep versions of Windows for different
> architectures in house? I'm not discounting the point made by Craig and
> a few others that Microsoft is perfectly capable of finding another
> partner if Intel got too 'uppity'.

Sure it is a risk that Microsoft could decide to abandon Intel and go
with a completely different architecture. However, is that risk greater
than IBM convincing other major Linux players to use the Power4/PPC
chips as the reference platform? I think you could agree that those
chips are as powerful or more powerful than any of Intel's offerings.
With IBM moving completely to Linux internally and using that as an
example of how other large organizations can do the same as well as
other countries moving to Linux (Asian countries primarily) I think
Intel has a LOT to lose IF they decided to simply toe the line with
Microsoft and NOT heavily influence Linux developers/vendors/etc to
continue to use Intel architecture.

> But they loose their near monopoly on the desktop and they actually have
> to compete for the server space on the technical merit of their processor.

I still don't see why Intel would instantly lose their monopoly because
they are supporting Linux? Microsoft, if they decided to change
architecture would either have to run all current code via emulation
(Which will most likely break quite a few apps not to mention be slower
than the hardware ia32) or convince all the 3rd party developers to
transition to the new platform, do necessary rewrites/testing/etc and
effectively support two hardware platforms for a transition period
(several years).. not to mention an NT based platform, Win9x platform,
etc..

Beyond that, whose chip would Microsoft use? IBMs? Suns? Would they
develop it in house (The total Microsoft lock-in?) -- Perhaps I am
missing an option, but at least to me, none of those sound that insanely
great .. sticking with a ia32 compatible chip seems to be almost a must.

Perhaps I am simply not seeing the big picture, but to get back to your
original question, I think Intel ultimately is risking more by NOT
taking a huge interest in keeping Linux primarily on Intel hardware.
Furthermore, I think Microsoft truly doesn't really have any great
alternatives than to simply stay with Intel at least for the next few years.

Joe