|Old rockers can keep
going for decades. Mick Jagger, Elton John and Paul McCartney have proved that
old timers can shake it as well as the leanest nouveau boy bands. But the old
reliables stay in the charts, not by endless re-releases of their classic
standards, but by working hard on new material. |
It's three years
since Sun Microsystems was at the top of the server charts. Since then the once
hot SPARC server maker has lost its business sparkle, and its lame comeback
attempts have seen a cool reception in the market. As the quarters roll by and
Sun's losses keep mounting up, I ask the questions?
Are Sun's days
Is there a comeback strategy for the old rock star?
what you'll read in the ill informed financial press, Sun's decline was not just
due to a single episode or failure. Just as its success was not to a single
achievement. Instead a lot of little downward steps added up to a great big
fall. I've been charting this downward progress here in the SPARC Product
Directory for many years. You can see what I think were the critical events or
non-events in the column on the right.
If you're a customer, or partner
or employee of Sun, you probably just want to know these simple points:- Is the
company doomed? Should you bail out now, before the rush sets in? Or is it worth
waiting for better times? And if better times are coming, how will you
recognise the signs towards that progress.
I believe that Sun has
several different possible futures depending on what it starts to do now.
#1 - the Same Old Song
If Sun carries on in the same old groove,
with the current separation of its SPARC and Linux strategies I think there's a
40% probability that Sun will go bust or be acquired within the next 18 months.
I've seen plenty of better performing companies than Sun head in that direction
and get into my acquired,
dead, renamed, merged & gone away companies list.
happens there will be a future for SPARC based systems, because Sun doesn't
actually make its own chips, and others, including Fujitsu, have designed SPARC
chips in the past. But apart from a simple design rule shrink which would give
another two year kick to performance, SPARC would cease being a commercially
competitive technology after another two years. Beyond that, it might still be
used for five years in military and embedded systems, but there would be no
future without a strong OS sponsor behind it.
Future #2 - Get the
Business Consultants in and Make Sun More Profitable
would cut investment in technology and dispose of business units which were
non-core. It's easy to imagine what the hatchet men would do in the case of
Exit the unprofitable Intel/Linux business. Sell off Java as a
medium sized software company. Chop away at the unprofitable entry level SPARC
server range. Sack 50% of the VARs and take more business directly in the mid to
high end SPARC server market. Stop piddling around making the world's most
expensive me-too network storage, and buy all of it in from outside.
would work fine for about a year, and Sun could get good profits that way on 30%
lower revenue and being a $5 Billion company. But what would happen after that?
Cutting back on technology would mean that the company could continue being
profitable for maybe another two year as the revenue trickled south to $1
Billion. In three years the company would have lost its edge, the SPARC
processor line would be regarded as a joke, and the company would have a slow
lingering death as a small services organization living off a customer base
which was just too lazy to unplug its legacy systems.
criticisms you may level against Sun's CEO, Scott McNealy - being parsimonious
with development budgets isn't one of them. Fortunately Sun hasn't started down
this route yet. But if the top management changes, it could still happen. So
Future #3 - Learn a New Song - Take a Gamble on SPARC/
Linux is the only Unix OS which has a credible long term
future. The Unix world got a slap in the face by Sun's antics with Solaris x86
in recent years. The "now you can have it", "now you can't"
games from Sun woke up even the doziest of techies asleep at the back of the
class, with a clear message that the operating system they had bought into
might have had open systems roots, but it was in fact a tightly controlled
proprietary OS. Single sourced with all the advantages and disadvantages in
terms of cost, risk of lock-in and availability that come with that proprietary
I've said it before, and I'll say it again. Sun needs to
commit to a transparent SPARC/Linux server line.
That will increase
confidence in the SPARC/Solaris user community and will be much more credible
than Sun's lame x86/Linux products, which most Linux buyers (over 99%) wouldn't
touch with a bargepole.
The only problem is that at the entry level end
of the SPARC/Linux server range, Sun would have to price SPARC servers below
those of higher powered Intel architecture boxes. That's something which the
company has always shied away from in the past.
implemented SPARC/ Linux strategy could grow Sun's revenues and take the
company back into profit.
On the other hand, if Sun does this too late,
or in a half hearted manner, the alternative futures I've mapped out above will
come to pass. And then all you'll hear about Sun in the future will those low
cost "best of" compilation CDs they advertise on TV at Christmas.
also:- Why Sun
Should Acquire a Solid State Disk Company (May 2004)
the mid 1990s Sun Microsystems was a cool band. They played sweet songs like "Open
Systems". They drove fast SPARC accelerated systems. And they made money
for their investors. It looked like they could do no wrong.|
Did Sun Microsystems Fall from Grace?|
It wasn't just that the songs
got less sweet, or the hubris or the (Java) drugs. Fashions are fickle. The
times changed, and the fans changed but the old rockers didn't notice or didn't
seem to care.
Sun's decline was charted in many articles in the SPARC
Product Directory as it happened. Here are some of the highlights.
1996 Sun stopped actively promoting its "SPARC" brand and
instead Sun and Java became the new brands. Later when most other SPARC server
companies had been driven out of business, this was taken as a sign that Sun may
be playing the open systems tunes, but its real tastes were proprietary.
1999 Sun's star shone brightly enough that it could have killed off the
fledgeling Linux market by launching its own range of Solaris x86 servers, and
promoting its OS as an open source standard. But Sun clearly gave the
impression that it didn't want to soil its hands with that fithy Intel hardware.
Four years later, when Sun tried to go down that route. It was already well worn
by others who had been there before.
In 2000 the trendy tunes
in the computer market were all about network storage. Sun tried to get into
that. But it had too much of a loner image to fit in with all those Intel
server users. And it was too fat to squeeze in as a low cost supplier. We said
it wouldn't work at the time. Sun spent hundreds of millions of dollars to
acquire a new image. But what the market saw was mutton dressed as lamb. It
In 2001 Sun got hit by a treble whammy. The
dotcom generation, its biggest fans, were getting old or had passed on. And
reliability problems dented the cool image of its SPARC servers. And actually
when you took a closer look at those SPARC processors they didn't seem so fast
any more. Sun had lost its edge.
In the next few years Sun's revenue
continued to decline. Its profits disappeared. It tried to make a lame comeback
by playing some newer Intel/Linux tunes. But if you looked closely at their
videos you could see the group wasn't really singing at all. They had been
dubbed. Sun had become a follower of fashion, and was no longer a leader. But
can Sun still make a comeback?
- 10 days after publishing this article yet another option emerged...|
and Fujitsu May Merge High End SPARC Server Product Lines
October 24, 2003 - a news story today in
speculates that Sun Microsystems may reduce development costs by pooling
resources with Fujitsu on the design of future high end SPARC servers.
wouldn't be the first time that Sun teamed with a SPARC partner in the mainframe
business. Sun and Cray Research codesigned the backplane and architecture of
Sun's barnstorming SPARCcenter 2,000 which was launched in 1992. Sun got the
kudos from that design, while Cray's own SuperServer 6400 garnered virtually no
sales due to poor marketing.
A decade ago there were dozens of
commercial SPARC mid range server companies, and as many as 6 companies which
made mainframe class SPARC systems (Solborne, Auspex, Cray, Meiko, ICL and Sun).
But today Fujitsu is the only mainframe company still using SPARC technology
apart from Sun.
An agreement between Sun and Fujitsu on high end SPARC
development would do more than save Sun hundreds of million dollars annually in
development costs. In recent years, Fujitsu's own PrimePower SPARC servers have
outperformed Sun in public benchmarks. Also Fujitsu's independently designed
SPARC servers did not suffer the ignominious reliability problems which hit Sun
a few years ago.
Fujitu's flagship products have often been better
than Sun's, but Sun's marketing has always been superior. In the late 1990s -
Fujitsu had to write off its failed investment in HAL Computer - which
developed US designed 64 bit SPARC chips, workstations and servers.
few days ago in the editorial office I was discussing the possibility that
Fujitsu may try to acquire Sun. We agreed it wouldn't work because it would be
like a hippo trying to manage a tiger. The hippo would have to sit on the tiger
to keep control of what it was up to. And having a tame tiger which can only
move at a snail's pace wouldn't be much fun. This new agreement, if it goes
ahead, will solve some, but by no means all of Sun's many deep rooted business
...Later - 8 months after publishing this article
this deal was confirmed publicly.
SANTA CLARA, Calif., and TOKYO -
June 2, 2004 - Sun Microsystems, Inc. and
Fujitsu Limited today announced the expansion of their 20-year
strategic relationship to include the joint development and delivery of future
generation Solaris and SPARC-based systems.