Inexplicably, we’ve gotten through much of 2006 without Linux completely kicking Unix out of the market. Analysts and Linux faithful are at a loss to explain how Sun Microsystems’ server revenue climbed almost 14 percent since the second quarter last year, pushing Sun ahead of Dell in the rankings. Gartner pegs Sun’s Unix server market share at 56.9 percent.
Playing analyst for a moment, I’ll send myself an e-mail explaining that having majority share of a dwindling market is nothing to crow about, and that a 10-foot rise from a mile-deep hole is hardly a rocket-like ascent. (I can be such a killjoy.)
The truth is, considerations of my alter ego and his brethren notwithstanding, Sun’s growth is a victory, and I will shelve my humility yet again to point out that I’ve been certain of Sun’s renewal since I started writing this column. It is a matter of patience, vision, and the nerve to stick with that vision even when nobody else seems to see it Sun’s way.
Sun’s vision is to make Solaris on Opteron an industry leader. Unlike IBM and HP, which treat Opteron like a second-source Xeon, Sun invested in unique engineering that leverages Opteron’s superior performance and power utilization characteristics while meeting enterprise expectations. Sun’s Opteron systems are not the vendor’s cheap seats reserved for those who can’t afford RISC. Sun is as serious about AMD64 as it is about Sparc.
Sun also stretched its neck out -- way out -- by daring to differ with the seemingly universal consensus that Linux is the path paved with gold. Is there a compelling reason to abandon Unix? Maybe the qualities that made Unix an enterprise mainstay still exist. Linux faithful mock Solaris with the adolescent nickname “Slowaris,” but I challenge critics to find one instance among the innumerable huge-scale Solaris deployments where decision makers are considering a switch to Linux for speed’s sake. The only intelligent argument in favor of taking Unix down was that it was closed source. So Sun took the bold step of open sourcing its crown jewel, Solaris, to take the “proprietary” millstone from around its neck. Now Solaris is the only one of the Big Three Unixes that is open source.
Sparc, too, has rallied. Sun has long maintained power-efficient processors in its product line, and the eight-core, 32-thread UltraSparc T1 trounces competitors with peak power utilization of 79 watts. Each core runs at a maximum of 1.2 GHz; but eight cores, four threads per core, RISC architecture, and on-chip memory and bus controllers give Sparc the advantage over dual-core Intel x86 CPUs in clock-for-clock comparisons.
What really puts Intel in the doghouse is Sun’s decision to open the design of UltraSparc T1 for public use. It is the silicon equivalent of open source, and it’s no lip service. Inexpensive and readily available programmable logic puts anyone a few hundred dollars away from being able to mint their own 64-bit Sparc CPUs. Of course, it takes more than that to make it do anything; but the point is that Sun has matched IBM in opening its server processor technology.
Sun’s growth is good news for the entire industry. It proves, as I keep pointing out, that trends are useless in predicting the future. Find players with vision, drive and patience, mixed with a desire to please customers as well as shareholders, and it’s easy to pick the winners. Sun’s a winner.