Perspective: Why Sun invited Intel to the party

I've been racking my brain, but I still can't find a conspiracy at the heart of the Sun-Intel alliance.

In case you missed it, Sun Microsystems agreed to put Intel processors into some of its servers, while Intel will help promote Sun's version of Solaris for servers running x86 chips. Sun has previously used Intel chips, but it phased them out.

When a big story strikes, it's our job in the reporting ranks to flesh out some sort of all-encompassing strategy--especially when two large, often antagonistic companies suddenly join hands and frolic to "So long, farewell, auf Wiedersehen, goodbye" in front of their investors and competitors.

Does Sun need Intel to alleviate supply concerns with Advanced Micro Devices or give it credibility with overseas banking clients? Is the real story here IBM? Does it open the door to an Intel-Sun-Apple launch into the home and IPTV? Haven't the two companies shared a long, antagonistic history?

And what on Earth are NetBeans?

In a way, playing the board game Risk as a kid prepared me for writing those kinds of stories. A single move on the board can upend a swarm of alliances and strategies. But the problem with conspiracy theories is the same problem with Risk. There are always a few factors that get a disproportionate amount of influence. Think about it: the goal in Risk is to take over the world. So how come the best way to win is to control New Guinea and Peru?

The reality of the Sun-Intel alliance is fairly simple.

The reality of the Sun-Intel alliance is fairly simple. Between 2003 and a good portion of 2006, AMD had a better server chip. Corporate customers wanted it, so Sun adopted it. Last year, Intel upped AMD with Woodcrest. It can now offer pretty good chip pricing and supplies because it got to 65-nanometer manufacturing a year ahead of AMD, and it has more factories.

Corporate customers have been adopting Sun's version of Solaris for x86 servers, so Intel will help promote it to see if it can help displace more of those lingering RISC-Unix boxes.

Performance, price and volume--that's pretty much the story.

While that sounds dull on the surface, it's actually a somewhat significant milestone. Sun is bringing Intel chips back into its fold because of convenience. That's a sign that the microprocessor market has matured. Server makers can choose AMD or Intel. AMD isn't fading away or stumbling drastically, something that has occurred in the past.

Then again, AMD has to deliver on performance, and that could be problematic as time goes on. AMD is a much smaller company than Intel and has many fewer engineers. Most of the time, more is better. Thus, AMD may find the gains it has made over the past three years receding slightly with each generation of chips.

More doesn't always equal better. AMD's credibility with corporate buyers began because it came out with a better chip, Opteron, while Intel continued to fumble with the Pentium 4. But you can't really count on that happening again soon: it was an unusual combination of circumstances that may not occur again. Of course, Intel likely will continue to experience erosion in chip prices.

The newfound dullness at Sun, meanwhile, is significant as well. In years past, Sun events weren't complete without a) a cartload of unintelligible code names and 2) lots of bombast. Sun would declare that it's superior to everyone else, that its competitors were engaged in a plot to take over the world, and that if you didn't understand it, you had the mental capacity of a gibbon.

The high point for that type of announcement came two years ago when then-CEO Scott McNealy talked about the battle of "humankind versus IBM Global Services, and we are kind of the leader of mankind in this aspect." IBM Global Services? It just seemed like such a B-league evil. If Spider-Man had to face it, he'd probably outsource.

By contrast, Sun's current CEO, Jonathan Schwartz, said he simply went out to dinner with his counterpart at Intel, Paul Otellini, and they hashed out their differences.

Biography
Michael Kanellos is editor at large at CNET News.com, where he covers hardware, research and development, start-ups and the tech industry overseas. He has worked as an attorney, travel writer and sidewalk hawker for a time share resort, among other occupations.

More Perspectives

[scdecade]

Fluff piece

You don't know what netbeans is?!? The real story is that there is no story? Huh? Why is the title "Why Sun Chose Intel over AMD" when Sun did not choose Intel over Amd? Did I miss something here? When did Sun say they were going to stop making AMD powered servers? How does this negatively impact Sun's relationship with AMD in any way?

There's 2 real stories here. First, Intel needs Sun more than Sun needs Intel. Recently Intel has been publicly saying that the way software is designed needs to change to take advantage of multicore processors. This is the exact same rhetoric we've heard from Sun for the last half decade. Second, one of the major global IT players has finally stood up and said what nobody has wanted to admit. Linux is not well suited to systems greater than 4 way. Partnering with Sun is a lot easier than wading into the Linux swamp to fundamentally re-architect the OS while attempting to maintain some semblence of backward compatibility. In cases where it's more difficult/expense to make a copy of your data than it is to buy a bigger computer Solaris is peerless and Linux is useless, or at least very kludgy.

There's your story sonny boy.

[YankeePoodle]

Multi-Core processors

You have a good point about the multi-core processors that are to be released by Intel (one prototype of upto 80 cores). SUN has been a proponent of this strategy, probably has the best operating system till date for that kind of environment.

I am surprised but I think Intel should get the pat in the back of the amazing turn around, after AMD has munched a significant portion of Server Processor market.

Sun is doing the right thing, with basic understanding "Deliver to the customer needs" rather than unproductive rhetoric of Scott McNealy.

I dont know if this strategy will pay off, but it sure looks loaded with potential.

Cheers!!!

[Waam]

Sun at work, Linux still at home.

This is not a slap in the face to anyone. The real story here is
that the companies are just doing what the customers want, and
that is Solaris on the fastest chips, Intel's new muti-core
chipsets. Thanks Sun for all the options, it's like somebody
turned on the light in the closet over there.

Intel benefits the most between the two companies because they
tap another company with furiously loyal customers and expand
their portfolio. Sun customers benefit the most overall because
their loyalty to Sun is finally not misplaced (you guys know what
I mean).

If I weren't hammering on this HP everyday, and had a choice to
start up from scratch. I would go back to Solaris. By now,
everybody knows Solaris is just better suited for multicore
processing than sitting back on my older systems here waiting
to see if new upgrades to multicore chips are going to actually
bust my current Linux set-up.

[Dachi]

I would be a terrible writer

Most people deploying x86 servers today are not interested in deploying the second best product.

If Sun does not meet demand for Core 2 based deployments someone else will.

[chuck_whealton]

Inviting Intel to the game

Sun should have invited Intel to the game just for competition sake.

Whereas the author appears to be counting AMD out of the game already, I wouldn't be so quick. AMD has been around for a while and has done some darned good work.

In fact, if I were the writer, I'd hope that AMD didn't just get lucky. If they did, I wouldn't count on very much more "pricing pressure" being exerted on Intel.

And we all know what that means.

Charles R. Whealton
Charles Whealton @ pleasedontspam.com