Bernstein Analyst Calls for Sun-Dell Partnership

by Timothy Prickett Morgan

One of the fun things about the computer business is thinking about the possibilities, and that is exactly what Toni Sacconaghi, the main IT analyst at Sanford Bernstein, has done in a recent research report. To amuse himself and Sanford's investment customers--and to get a little PR like this piece in this newsletter--Sacconaghi is proposing that we ponder "the unthinkable." That would be a partnership between Sun Microsystems and Dell.

When you stop laughing or shaking your head, read on.

Sacconaghi says that he has no reason to believe that any partnership talks are underway, but he says that some kind of deal akin to the one that Dell has with disk array maker EMC might be beneficial to both Sun and Dell. This might sound crazy, but Sacconaghi makes a good point when he says that Dell is losing market share in the high-end of the server market, where servers cost more and have more processors. With the soon-to-be-announced Opteron-based "Galaxy" servers from Sun, Dell could, says Sacconaghi, flesh out a unique high-end server offering that can run Windows, Linux, or Solaris. Considering that Dell's top brass has just nixed the idea of using Opterons in the PowerEdge server line after mulling it over, the idea of Dell selling Galaxy Opteron boxes seems pretty remote.

Many people were calling for a Sun-Fujitsu-Siemens triple-play partnership on Sparc iron for many years, and Sun finally threw in the towel on its "Millennium" UltraSparc-V designs and went with tweaked future PrimePower servers based on Fujitsu's Sparc64 clone of the Sparc chips for big SMP boxes for running monolithic workloads; Sun is still developing its "Niagara" and "Rock" multithreaded chips, and will eventually try to convince customers that these are even better than the Sparc64 boxes it will sell for the next few years as a stopgap. This deal took years to hammer out, and that was between three partners (well, in many respects, it is really between two partners, Sun and Fujitsu).

Sun and Dell are enemies, and there are not enough years left in the 21st century for them to come to a deal, unless something really bad happens to either company. Sun might consider having a big distribution partner for its Galaxy machines, but what Sacconaghi doesn't understand deeply enough (or ignored for the sake of some PR) is that Sun doesn't want to partner with Dell to push Galaxy, but rather that it wants to use Galaxy running Solaris 10 to hurt Dell, and hurt it badly as Dell tries to push Linux and Windows on what will arguably be inferior X86 iron.

What Sun and Dell should do is partner to make sure that Solaris 10 is an option on the PowerEdge machines, but considering that Sun will try to compete against Dell with the mildly superior performance and technical characteristics of Solaris 10 compared to Linux 2.6, this seems very unlikely. Dell will surely install Solaris 10 on any PowerEdge server that customers want--and it will do it today. But Dell is not going to go out of its way to help Solaris 10 on X86 get more traction, not when it is trying to promote either Windows or Linux as the answer to everyone's data center problems. If Sun was heading toward bankruptcy, Dell might acquire it or Sun might concede to partner with Dell in the manner that EMC has. Other than that, these two will remain bitter rivals.