Can Sun get posh with Bechs?
Time to end the waiting game
Always betting on the future
A big part of the Sun's strategy continues to revolve around products that could pay off in the future.
A rather frank McNealy admitted that Sun realizes the vision talk is vision and that the company doesn't expect customers to move en masse to grids, thin clients, free software, open source chip architectures and the rest of what Sun calls its "participation age" program right now. Hearing this confession helped reassure us a bit. Sun often gives off the impression that current revenue relies on its vision of the future, and that seems a rather untenable position for a public company to hold.
To its credit, Sun lays out one of the more acceptable plans for moving technology forward - one we think most of the industry will adopt. Companies will not build or manage their own data centers in the future. Billions of internet connected devices will flood the network with all kinds of information. And - dare we say it - the thin client style of computing will eventually be the dominant technology for businesses and possibly consumers in poorer countries.
Sun insists that it's the only company that will be able to deliver all the back-, middle- and front-end technology for this grand future.
"It won't be IBM, HP or Dell," said CTO Papadopoulos.
We, however, suspect IBM, HP and Dell will have some role to play in computing's future. Call it a hunch.
Sun's real point is that it's the only infrastructure vendor spending billions on this vision right now.
The research and development spend at Sun has long horrified the brainiacs - and we use that term mockingly - on Wall Street. We happen to find Sun's penchant for experimentation endearing - as long as a payoff does arrive at some point. So, keep at it, boys.
The Bechtolsheim Effect
We continue to think analysts underestimate the effect Bechtolsheim's team will have on the x86 server market. Bechtolsheim has the most glorious track record imaginable from an engineering and investment standpoint. This man knows where markets are heading - see the original SUN workstation, Granite, Kealia, Google and VMware.
Sun has "dozens" of Opteron-based systems in the works, including an 8-socket system already shipping to select customers, blade servers and the Thumper dynamo.
"For the first time in years, there is real excitement in the industry standard server space," Bechtolsheim said. "I know the Intel camp is working overtime. I am not trying to pick on them. (But) the reason we are doing Opteron is because it's faster, it consumes less power and it's more scalable."
"In reality, we haven't seen one benchmark running faster on Intel than Opteron."
HP and IBM have made a decent Opteron investment, but they didn't have a chance to start from scratch the way Sun did. They have vast Xeon businesses to protect and must treat their Intel defection with caution. By contrast, Sun can repeat what it did 24 years ago (hence the birthday cake) and try and change the game.
In a perfect world, Bechtolsheim's fleet of Opteron boxes would have arrived about six months ago. Sun seems forever trapped by "huge company" delays when one might expect it to move a bit quicker than giants like IBM and HP. All that said, we have to believe Bechtolsheim will deliver some shockers this year.
Sun will only start generating serious profits if this Bechtolsheim gear sells well. The kit is key to Sun's immediate ambitions and to its goal of putting a thin client and OpenSolaris in every mud hut in Africa.
It's Bechtolsheim or bust. ®