Original URL: http://www.theregister.co.uk/2009/04/23/sun_no_kidding/

Sun: Let us now praise the ponytail

Could you have done better than McNealy and Schwartz?

By Andrew Orlowski

Posted in Servers, 23rd April 2009 00:40 GMT

The end of Sun Microsystems as an independent systems company has been met with a mixture of "I told you so" and "good riddance". But as someone who spent plenty of time at Sun over the past decade - including the dot com years - I wonder if anyone could have played a better hand?

Schwartz and McNealy succeeded in stemming the haemorrhage of Sun's user base, preserving most of the company's R&D investment, and ultimately delivering a handsome deal to shareholders. The latter is what CEOs are actually paid to do - not clean your car, or cure malaria.

(Or as our Ashlee put it here, "replace data centers with jellybeans and justice").

Giving McSchwartz credit isn't always easy. Would a grown man, in full possession of his marbles, really write "If the Information Age was passive, the Participation Age is active" and give the blog post the name "inevitability"? Hubris beckons. For any reporter, it was hard to keep up with Sun's desktop strategy. In a mere three years, it shifted from emphasising the thin-client SunRays, to its own Linux ("Java Desktop System"), to giving away a souped-up Solaris for x86.

(In 2002, Sun had axed Solaris for Intel and it took pressure from customers to revive it. By 2005, it was a key strategic asset again, and today, you can buy Dell Poweredge servers with Solaris).

My esteemed colleague Chris Mellor thinks Sun could never decide which company it wanted to be - a proprietary SPARC vendor or a commodity Intel vendor. But that dilemma has faced not only Sun, but every proprietary hardware company, for every day since the ascendence of Intel in the early 1990s. (Remember that even Microsoft didn't bet the farm on Intel, making sure NT ran on MIPS, Alpha and PowerPC chips, all of which were superior at the time, and looked like they would endure). It's really the same challenge Sun posed to minicomputer makers in the 1980s with Unix. Most of those competitors were sold for scrap, not $7.4bn. The mighty SGI fetched just $25m recently.

It's about the hardware

Sun's fortunes have fundamentally depended on making sure its hardware customers keep coming back. If you think about it, that's actually a lot easier for a Sun, with an established base, than a Dell, or your corner store PC assembler. The hard part is ensuring the price/performance of the kit is competitive, so for a CIO, leaving isn't more expensive than staying where you are. And on this criterion, Sun's historical performance isn't so bad, once you strip out the anomaly of the dot com years, when companies splurged on Sun kit.

Revenue in the most recent full financial year was £13.8bn, with a positive net income of $403m. In 2003, revenue was $11.4bn, with a net loss of $3.4bn. The McSchwartz years also grew Sun's services income from $3.6bn to $5.2bn. That's not too shabby for a couple of guys who apparently don't know what they're doing.

Sun's success in the internet bubble really was a freak episode. Sun was a smart workstation vendor who found itself handed the Cray's 64-bit E10000 from SGI. For regulatory reasons, the US Gov didn't want one company (SGI) owning too much supercomputer gear, it was asked to divest, and Sun was the beneficiary. But sometimes you earn your good luck, and the E10K years allowed Sun to establish a beachhead in datacenters and escape the workstation business before it finally collapsed.

The giveaway software strategy (Java, Solaris, OpenOffice) proved smarter than the Kumbaya hype that accompanied it. It didn't cost much, and it was Sun's best hope for acquiring new customers. Solaris is quality from top to bottom, and luring people in with freebies at least gave the sales guys a chance to get a foot in the door.

But Sun's fortunes - and its value as an Oracle asset - depended entirely on hitting the price/performance. It was all about hardware. And here, Sun had to cope with some serious failures. The seven year death march of UltraSPARC V (Millenium) chip (canned four years into the new millennium without ever seeing combat), Gemini, and a Rock variant of Niagara that's years late. Opteron offered only a short-lived advantage, and it fell to Fujitsu to mastermind the big iron.

Yet by betting on some clever systems thinking during the down years, and backing their R&D departments to come up with the goods, McSchwartz ensured such disasters were survivable. The result is a lot of in house expertise in virtualisation and threading, that can make a difference with real workloads.

If Oracle isn't completely dumb, it will appreciate quite what an incredible asset it has acquired - because this know-how can help every part of its business. Peoplesoft cost Larry $10.3bn. Sun looks like a bargain. ®