I REMEMBER WHEN.I remember the first time I saw an SGI. It was at a
conference in DC, about ten years ago. Something quite new,
they stood out from the crowd, and left an impression. Over
the years this impression grew. I have ported countless
software to the various Unix platforms over the years,
SunOS, Solaris, SGI's IRIX, DEC OSF/1 and Digital Unix,
HP-UX, AIX, and countless others, most of which are
deservedly forgotten. This is the aspect of their business
with which I am most familiar, although I am aware of their
excellence in graphics, media, and countless little seeds
they have planted.
Now, SGI has always been good at creating positive
impressions. Taking an SGI out of the box and setting it up
is always so much more exciting then their peers. Colorful
machines, multimedia in the firmware, little video cameras,
fun little applications, always entertaining. But that just
speaks of their nature. And that is a valuable thing, to be
unique, to be where the others aren't. I like that. But
I'm really interested in what's on the inside.
Here is what I learned. SGI knows Unix. Of all the other
flavors of Unix, SGI has the best compatibility, the best
development environment, and the most knowledgeable
engineers. The machines perform well, rarely break, and
compatibility from release to release was relatively smooth.
I always wondered why Digital, IBM, and HP could never
achieve this with their vast resources.
Now, this experience was with IRIX and MIPS. Before Linux
was "in". Before SGI's lame Microsoft strategy. Now SGI is
leveraging their Unix expertise, their ability to scale with
multiple processors, with their two of their key strategies,
the Linux server market, and their new flexible servers
built from bricks. This is a key part of their future. SGI
knows Linux. They have made the most contributions to the
Linux community, and are constantly improving their own code
base to keep themselves at the leading edge in performance
and usability. This is necessary to differentiate oneself
from a cheap PC running Red Hat, which may serve a small
business, and cannot handle any serious computing needs.
The simple fact is, the competition has provided SGI with
the opportunity. Only Sun has really been able to hit the
nail on the head. IBM does not not what they are doing.
They may have good sales, and the myth of Big Blue, but they
make life difficult for software developers. A recipe for
technical failure, but then again, they dont really depend
on others to make it, so they can screw the world and
survive. HP is another disaster. Stubborn fools, I'm
sorry. Can't fix a simple bug even when you hold their hand
and point them to the code. And forget compatibility.
NT and Linux. Well, you get what you pay for. I would
rather work on a low end SPARC than the most powerful PC.
Once you start using quality components, like keyboards that
dont wobble, SCSI disks, and nice monitors, gee, the prices
aren't that different anymore, are they? And Microsoft makes
developers lives as miserably as possible. They do have
good compilers and tools, but every other one of their
products suffers from poor ergonomics, unstability, and bad
performance. The X Window metaphor is so far superior, and
its general UI is incomparably better. Not to mention the
agony of getting your application to work on NT, especially
if you need to have a professional product, which must run
on multiple platforms.
As for Linux, well, SGI has the best Linux solution. Real
machines tuned for multiprocessor performance. And they
have a small stake in VA Linux, which has opportunities in
the low end. People should not balk at spending more money
to get a product which is going to perform better, work out
of the box, and backed by a team of experts who understand
your business. SGI can give you a Linux solution with
hundreds of processors that rival the Cray of just a few
years past. Only more flexible, more compatible, based on a
plethora of industry standards.
Yes SGI's past has been colorful. The industry darling in
their infancy. The first choice in graphics. Successes and
failures. Now lets look at the numbers. Revenues, 2.3
billion. Almost triple Yahoo. Return on equity -60%. Nine
times better than Amazon. Gross margins, 35%, in line with
their industry. Price/Book .65. Five times better than
Apple, 9 times HP, 18 times IBM. Shareholder experience.
Well, funny, you would think that shareholders are suffering
tremendously, but, hey, amazingly enough, if you made steady
purchases of SGI shares over the past two years, you are
probably only a little in the red.
But the numbers are only a picture. A picture of a random
piece of the past. You have to look at the future. And for
this you have to understand what seeds have been planted.
Technical expertise in Unix and media. A culture of
independence, yet the ability to cooperate. And the
partners and customers they have nurtured. The respect they
have earned. A broad range of servers, a broad range of
graphics and media solutions, including two subsidiaries,
Alias/Wavefront, and the recently created Kasenna. Stakes
in VA Linux and WAM!Net (to whom they also lease part of
former Cray Headquarters). Inventiveness about. The ability
to deal. These are the things to which I would assign the
greatest weight. |
