Sun can run, but can it hide?
By David Berlind, Enterprise
May 7, 2002 2:30 PM PT
URL: techupdate.zdnet.com
After interviewing Intel CEO Craig Barrett, you can't help wondering about Sun's current commitment to its SPARC processor lineup. Either Sun CEO Scott McNealy knows something about microprocessors and systems that Barrett doesn't, or Sun is in deep doo-doo if it doesn't do something fast.
Based on Intel's experience in this area, the former is highly unlikely. So what does that leave you with?
The way Barrett sees things, if you're a microprocessor manufacturer, the price/performance ratios of your offerings are your only differentiators. What about the other important stuff like reliability? According to Barrett, those are not differentiating factors. Either you're reliable, or you're not. If you're not, then you lose your seat at the table. But once you have that seat, price/performance rules.
This, of course, is nothing new. What is new is that Intel has a seat at the 64-bit table with 64-bit veterans Sun, IBM, Hewlett-Packard, and Compaq. Now that it's there, Intel is looking to apply the same price/performance formula that won it the top spot in desktop computing to 64-bit server computing. If early adoption is any indication of Intel's future in this category, then Sun has some serious problems ahead. Why? Of the four players that Intel joins at the 64-bit table, Sun is the only one that is not a customer of Intel's 64-bit products.
What's worse for Sun (and better for Intel) is that the merging HP and Compaq are ditching all their RISC architectures in favor of Intel's IA-64. The executives in charge of those architectures apparently saw the writing on the wall. IBM, the only vendor that can afford to drive Unix and Linux with two-processor architectures, will stay the course with two 64-bit offerings (IA-64 and its own Power).
Of course, those won't be the only companies selling IA-64-based products. We haven't even taken into account high-volume Intel pure plays such as Dell and Gateway that can move an extraordinary number of systems through direct channels. I'm not even going to bother discussing the market dynamics created when a whole slew of vendors competes to sell Intel-based systems.
Needless to say, Intel has proved time and time again that once one of its architectures hits its stride, it isn't long before that architecture dominates. It happened with IA-16. It did so again with IA-32. With its StrongArm architecture serving as the exclusive foundation for the two predominant handheld OSes (Palm and Pocket PC), Intel is on track to dominate the PDA space as well. When you consider all the partners that will be moving IA-64 for Intel, it's hard to imagine history not repeating again. Between its partners and the capacity of its fabs, it's easy to see how Intel can achieve the volumes that quickly lead it to dominating any market in which it chooses to play.
Which brings us back to Sun. The majority of Sun's revenues are derived from the high margins it's getting off its SPARC processor-based servers. Barrett, who knows this business as well as anybody, says McNealy can't hide behind those margins for very long. When IA-64 hits stride, and it will, Sun will need to do some soul-searching.
Perhaps it's time for Sun, and its new executive in charge of software, Jonathan Schwartz, to do some of that soul-searching now. Instead of sticking it out with a SPARC-only strategy, perhaps now is a good time for Sun to seriously consider giving its current and future Solaris customers the option of a seamless move to Intel. This would be similar to what HP is doing for PA-RISC customers and what Compaq is doing for Alpha customers. If Sun doesn't, it could risk losing those customers altogether. That wouldn't bode well for Sun if, at the same time, Intel starts beating it out of any new accounts.
Unfortunately for Sun, that's already happening. Barrett says Intel is way ahead of the goals it set for itself in terms of design wins--IA-64 victories in large commercial environments such as those found on Wall Street. It's not uncommon to see Linux, which Intel has worked to optimize for IA-64, show up in those design wins as well.
If Sun's server business does implode, it's not like the company has much to fall back on. AMR Research just released a report on the application server market that concluded that "as the application server market matures, there will be only four significant players remaining: BEA Systems, IBM, Microsoft, and Oracle." In other words, AMR isn't predicting a rosy future for Sun and its iPlanet application server. Sun is by no means the leader in the development tool or storage market either. While things look good for Java and OpenOffice, Sun is having difficulty capitalizing on either as a major source of revenue that could sustain the company.
So, perhaps it's time for Sun to come out from the server margins it's currently hiding behind. It can't hide for much longer. Says Barrett about McNealy's willingness to consider such a move: "If he wants and he's serious about a design win with us, we're willing to talk to him."
Do you think Sun should stick to its SPARC-only strategy or help its Solaris customers move to Intel's architecture? Why or why not? E-mail David or TalkBack below.
Stay focused: Sign up for Tech Update Today, the daily e-mail newsletter for those who need to know. |
