View from the Hill

Everyone's bashing Sun. So will SPARC survive?

It's become fashionable to knock Sun in the financial press, and even Business Week this week dedicated a long article to Sun bashing called "Will Sun Rise Again?" which includes another scary quote from Scott McNealy in which he says "...A computer is not a commodity."

If you still had any doubts about the thinking that's now driving Sun, that might make you more worried. But let's do a reality check here.. It was only a sound bite to a journalist. If you look back at Sun's history, their past success owes much to commoditizing products.

For example in the mid 1980's in the workstation market they used the low cost Unix operating system to beat out Apollo's proprietary OS. In the early 1990's Sun changed the multiprocessor server market from one in which only the government and oil companies could afford to buy supercomputers, into one where anyone could get a 4 processor SPARCstation delivered to their desktop from a mail order catalog. And then in the late 1990's Sun served the fledgeling dotcom industry by delivering mainframes which were preconfigured out of the box to run their ebusiness, even if those startup founders, which were drowning in VC dotcash, didn't know exactly what sort of business they were really in. So Sun has turned computers into commodities many times before and will do so many times again. You have to separate the sentiments behind McNealy's quote, from the reality. Sun would really like to charge more for their systems, but strong competition makes that unrealistic. So Sun has got to learn new tricks, or sell stuff cheaper like everyone else.

Another suggestion in the BW article was that Sun should maybe get rid of SPARC. That's another scary idea for most readers of the SPARC Product Directory but it's not new. I've been wondering about when Sun would do that ever ever since they launched their first SPARC systems in 1987. But burning your bridges without planning an escape route would be suicide for the company. And as we're all too well aware, Sun has for many years done too little to engineer a realistic escape route into Intel architecture servers. I discussed this issue in these pages in 1999 when there was less competition in the Linux market. But although Sun does have a small bridgehead into the Intel market, it's pathetically small and frankly if their Linux business was a separate company you wouldn't want to invest in it, even buying Sun shares would look like a better prospect.

So.. How worried should you be that your investmet in SPARC and Solaris might suddenly not look like such a good idea after all?

If you look at the worst case, then you'll probably agree that the prospects aren't so bad. Here are some ideas to dwell upon.

• Sun will survive McNealy's Departure. Lots of other companies have survived the unthinkable break with their founders as they got older, and markets changed. Hey, you never know! If Scott left Sun now (which would increase the share price by the way) then like Steve Jobs at Apple, and Arnold Schwarzenegger he might come back.

• SPARC will survive Sun. If Sun is acquired by IBM or HP, then the market for SPARC srevers will continue for at least another five years. When Compaq acquired DEC, they still carried on selling Alpha servers for a while, and the Alpha products were duds with a miniscule installed base. The installed base for SPARC is huge. Sun's revenue may be declining, but the number of SPARC servers has been increasing - even during the worst years of the recession. Sun doesn't even make its SPARC chips. And there's a huge invisible embedded SPARC market in telecoms and appliances in which Sun isn't even the main player.

• Apple has survived for 20 years in a PC market which they created and then gave away to Intel and Microsoft. Even if Sun shrinks to a $10 Billion company next year, it will still be twice the size of Apple, and unlike Apple, Sun still has partner OEMs which also sell SPARC into other markets which Sun does not address so well directly. So SPARC can survive and thrive as a 64 bit niche market, even if the rest of the market wants to stay in the 32 bit world.

That's all looking at worst case scenarios. The reality will not be so bad. Sun could get better at managing its business. The IT market could pick up. Someone might find a fatal bug in the next generation of Intel processors - worse than Intel's floating point gaff, and worse than Sun's cache problem. That would create a torrent of users switching to SPARC servers. A software company could create a new killer app for high security 64 bit computing which would make SPARC a necessity rather than a luxury. Sun could introduce a 32 bit SPARC appliance that sizzles at 3GHz.

There's going to be a future for SPARC for a long long time whatever they say in the financial press. Trust your instincts on this one. It's not going to go away. ...Sun Microsystems profile