Back in January, we considered whether Sun was doing the right thing by continuing in both the hardware and software markets. The verdict was a resounding "yes" with significant revenues in both the hardware and software markets, and a hefty chunk of cash rolling in thanks to support and services, it's clearly doing the right thing by sticking in those markets.
But can Sun really continue to compete? And if so, is it doomed to a future of playing second (or third, or fourth...) fiddle to the likes of Dell and HP (in the hardware market) and to Microsoft and even Red Hat (in the software market). Is it the Apple Computer of the server world?
On the hardware front, I think we'd all agree that Sun isn't about to become a desktop computer company. OK, there's still a market for high-powered graphics workstations, but this is a small niche and the alternative platforms (Intel-based PCs and PowerPC-based Macs) are snapping at the heels, performance-wise, of high-power Unix workstations. Sun's a server company, and this is where they're staying.
At the high end, Sun still does it right. The traditional competitors (most notably IBM and HP) are still there, and we're unlikely to see any new competition in this market as it's just too costly to attempt to get a foot in the door. So just as Apple has retained its traditional market in desktop publishing, Sun's unlikely to stop selling medium-sized and big servers.
At the entry level, though, Sun has had to think fast. Not so long ago it had decided that it wouldn't bother preserving the 80x86 version of its Solaris operating system, but as x86-based Linux machines became more and more popular as corporate email servers, hosted Web servers and the like, Sun had to do something if it wasn't to lose the market entirely.
Step one, then, was to dump the idea of being a one-platform pony and adopt the Intel architecture as its entry level: the acquisition of Cobalt the most popular Web appliance vendor at the time was an easy way to take a first step into Intel-based servers. Although the (now discontinued) Cobalt kit ran Linux, it was only a matter of time before Sun resurrected its x86 version of Solaris: although its x86-based machines can run Linux, they'll also sell you them with Solaris 9.0.
Sun's motivation for moving into the entry-level server market was fairly simple: as computers became more powerful, fewer and fewer people needed big servers, so they bought small ones instead. Sun didn't sell small servers and so they had to start doing so. Apple has responded to the same factors by moving the other way, though: as more and more people realised that the average desktop computer was becoming powerful enough to act as a Web server, so Apple figured that it ought to start shipping their kit in a rack-mountable form that looked and behaved like a server, with the option of some hot-swap bits for resilience.
Where Sun hasn't followed Apple, though, is in the latter's willingness to vanish down a completely unrelated tangent into the personal entertainment market. Apple's foray into portable music devices the iPod has been phenomenally successful (the best part of a million units were shipped last year) but stepping outside one's core competence is a risky business and not one that Sun is likely to follow.
It's fair to say, though, that Sun is pretty much the Apple of its market. Apple has its niches in desktop publishing and multimedia, but it has to look over its shoulder as others catch up; similarly, Sun is in the leading pack for the niche that is the market for big servers, but it has to keep on running. For entry-level hardware, though, it's as unlikely as Apple to catch up Dell or HP even though, unlike the Mac brigade, it uses the same architecture. On the software front, Solaris is very much a minority operating system - and even though the Intel version has been resurrected as a going concern, its dire hardware support renders it an also-ran.
There is a bit of a silver lining, though: Java. Here is something that Sun can point at, beam widely, and say: "We did that". The world feels good about Java, large numbers of software vendors exploit its superb cross-platform capabilities and millions of machines around the world have a Java Virtual Machine.
It's just a shame Sun doesn't make any money from it, really.