Sun and the great leap forward

When some look at Sun, they see a stumbling has-been surrounded by aggressive, hungrier competitors. Others see a technology visionary willing to take on Microsoft and Intel alone.

Me, I see Rick Bullotta.

Rick was my college roommate, and a fairly erratic one at that. An engineering major at Cornell University, he spent an inordinate amount of time spinning out his Ford Pinto in an icy parking lot. During one semester, he took no notes during any of his classes and adopted a Southern accent (the perfect accessory for his genetically attached 32-ounce beer mug).

At the same time, he was brilliant. He scored toward the top of his class in a competitive environment. When he got bored during tests, he'd take a friend's test blue book and fill in the answers for him. Currently, he's founder and chief technology officer of Lighthammer, a manufacturing software developer that can count Whirlpool, Matsushita and Bethlehem Steel as customers. (The name was chosen because it had both Tolkien and Zeppelin overtones.)

In the same way, Sun has thrived on goofy genius. CEO Scott McNealy likes to portray himself as a forward-looking leader who would, if he weren't saving humanity from Microsoft licensing agreements, be taking a softball team out for drinks.

He's also one of the few CEOs that will poke fun at himself. Asked last week about Sun's acquisition of Diba in 1997, he said: "About four minutes into the acquisition we said, 'Whoops!'"

Despite the lack of corporate formality, Sun has managed to come up with creative, unexpected solutions to difficult problems that often outshone the answers from larger, richer rivals.

Its UltraSparc chip has often been criticized as a performance laggard, but customers bought Sun computers in droves because of the harmonious relationship Sun achieved between its software and hardware.

This past week, the Santa Clara, Calif.-based company stated that it will come out with a processor with two cores this year. It will also release another, based on an entirely new architecture code-named Niagara, that will come with eight processor cores able to handle four instructions each.

"They have the most ambitious processor road map I've ever seen in my life," said analyst Peter Glaskowsky, editor of The Microprocessor Report.

Sun alumni have gone on to run Google (Eric Schmidt), Autodesk (Carol Bartz) and Transmeta (Dave Ditzel).

The company will use this technology to take on IBM and Hewlett- Packard with its N1 and Throughput Computing efforts. Under these strategies, new layers of software will be designed to coordinate the computing and storage power of thousands of systems, which ultimately will make it easier for companies or universities, for example, to offer faster, deeper access to information on bank accounts or libraries.

Sun alumni have gone on to run Google (Eric Schmidt), Autodesk (Carol Bartz) and Transmeta (Dave Ditzel).

The company, though, is dwarfed in terms of resources and money. IBM reported $81.1 billion in revenue and $3.6 billion in net income for 2002. Big Blue spent $4.8 billion on R&D and employed close to 300,000. Intel pulled in $26.2 billion in revenue and $3.1 billion in net income while employing over 80,000 in the same year.

By comparison, Sun pulled in $12.5 billion in revenue and lost money while employing fewer than 40,000 in its last fiscal year. Similarly positioned competitors--SGI, Data General--have tried to take on the giants in this way and failed.

Granted, many things the company tries do bomb. The JavaStation still ranks as one of the largest disasters in technology history, and Sun's efforts to sell Linux-based PCs will likely meet a similar fate. And even if Sun eventually persuades customers to switch to Linux, it's hard to see how the company will make money off it, even indirectly by hurting Microsoft.

Sun also has an annoying habit of applying multiple code names to amorphously defined technology platforms and experiments. Remember the early Java days? Initially, Java was to

Mathematically, Sun should not be able to survive. But discounting Sun's ability to mount a comeback might be a historical mistake.

change the entire computing environment; now, it's most prominently used to play games on cell phones. ="246433">Jini--whatever that was--was even worse. Do you go with the curly shoes and turban "genie" metaphor for branding, or a sort of Jimi Hendrix motif?

The MAJC chip was initially described as a multimedia processor that would likely cost around $20. Years later, Sun admitted it was a candidate to replace its UltraSparc server chip.

And let's not forget, large customers have been converting to Intel-based and Linux-based systems. Market share and profits are down. The company has been shelled in the past year by the telecom collapse, customer defections, and renewed attacks and improved performance by computing giant rivals.

Mathematically, Sun should not be able to survive. But discounting Sun's ability to mount a comeback might be a historical mistake.