Centerfield

Sun Microsystems does the Darth Maul

This is a bit of an experiment, as this post is about a big computer company rather than politics. Let me know if you think this is a waste of space, and I won't repeat it.

Sun Microsystems is doomed. After today, I believe that in ten years, it will have been sold for its assets like Digital was a decade ago. The remains of Sun may even prove just as toxic as DEC was to Compaq.

For those who don't know what Sun is, they sell workstations. Twenty years ago, they were the cool new business, newly enabled by the cool new 32-bit microprocessors of the time (Motorola 68k series was what they used). I was in heaven learning Unix and X on their 68010 and 68020 boxen; at one point, I even had my own Sun 3/60 (20MHz MC68020, 8M mem, 144M disk). Sun was deeply innovative from the moment of its founding to, well, when they pursued litation instead of innovation.

But one key to their success is that the PCs of the time were no threat. PCs were slower, and had much less memory and diskspace; they couldn't run anything much more powerful than original DOS, while Sun workstations ran Unix; in fact, all succesful workstations ran much better operating systems than DOS. But that difference has vanished. You can run Unix easily on the cheapest new PCs, in fact, more easily than Windows; for that matter, Microsoft finally has caught up to many of the things Unix workstations had.

Therefore, there's little point to paying extra to buy a workstation at this point. You can easily get a much faster Linux box than Sun workstation, paying less for the Linux box than you do to Sun. Sun still has staying power because there are customers that are still prefer compatibility to cost-effectiveness, and Sun has climbed the server value chain to the high end. But these strategies are both temporary staying measures. PC architectures have powered more cost-efficient and completely practical high-end servers for five years or so.

The reason they did Java was to try and move to a new product area, like Apple did with iTunes recently. Of course, Java was a limited success in the end; it delays, but doesn't end, their problem. Sun still needs to find some kind of business plan with long-term promise if they want to be around much longer.

I read an interesting article on their newest business plan. Here's a summary of what he says the strategy is:

Step No. 1: Make the argument that Linux equals Red Hat. ... Sun's view is that Linux is nothing more than Red Hat. The operating system is not about world peace and the charitable work of the world's great programmers. It's like every other operating system ever created: It's about the foibles, greed, mistakes and engineering prowess (or lack thereof) of one vendor -- in this case, Red Hat.

Step No. 2: Belittle Red Hat. By collapsing Linux into Red Hat, Sun now has a clear target. ... according to Sun, Red Hat is a very vulnerable target -- a company with limited resources, engineering talent, world coverage and capabilities -- with potentially serious intellectual-property issues.

Sometimes FUD works. Even so, this is so not going to work because Linux ISN'T Red Hat, and it's increasingly clear to people outside Sun. Hell, even if they succeed in making Red Hat look bad, that would just be good news for SuSe and Mandrake and the tons of other Linux distros out there. Sun will still be dead man walking. Red Hat provides a minority of Linux resources. But this will hurt Sun by distorting their view of their competition.

Step No. 4: Play up the OS-plus-platform advantage. ... The company is saying that you cannot be a legitimate, long-term player without controlling and harmonising the operating system and the platform.

...everybody knows how horribly Microsoft, Dell and Intel have turned out. This FUD won't make it beyond the press releases.

Step No. 5: Disrupt the market with a new pricing model. Sun wants its server pricing to mirror mobile phone pricing. When you buy a cell phone, you do two things: One, buy the operating system and the phone together; and two, subscribe to mobile services for monthly fees.

Remember, when you copy Red Hat, don't forget to say you're innovating.

Step No. 6: Feature customer choice. Sun ... is offering choice at the high end and at the low end. It is offering not only Solaris but also Linux and Windows for the operating system. And it is finally offering x86 via the Opteron chip from Advanced Micro Devices.

The only good point here, but it won't save them on its own - they need to give people a reason to buy Sun long-term, and the rest of the biz plan is sheer denial on that.

Step No. 7: Feature engineering. Sun is playing an old game here, too: "My tech is better than yours." It is saying that it will out-engineer not only at the operating-system level with Solaris but also on the hardware front. ...

They lost on the hardware front a decade ago. Now they're in crisis; their next-gen high-performance processor design was so buggy they had to can it. They're betting the farm on a somewhat experimental hyperthreading design that's all they have left. The Intel box on my desk already has hyperthreading; Intel's hyperthreading probably isn't as good as Sun's, but Intel is so much faster on single-thread perf, that I can't see how these will be overall faster for anything notable.

Sun still holds a slight advantage on the software front, but it fades with each year. I can't see it lasting before the kinds of resources that go into Linux.

On the x86 front, Sun is saying that Opteron -- AMD's answer to Intel's Itanium -- is superior to what Intel has to offer and that, through its long-term engineering experience, it is going to produce x86 products superior to those of Dell, HP and IBM. ...

AMD Opterons are so good that HP and numerous other vendors're already selling 'em. No Sun advantage here! Worse, it's very hard to get a big advantage in the PC market. It just doesn't happen, especially to giants like Sun, at least not without a much better plan. If they think Dell or other high-end PC vendors kinda slap things together thoughtlessly, they'll fare even worse than Compaq did.

Step No. 8: Feature the Microsoft-Sun deal. The money flowing from Microsoft to Sun will help. But more importantly, watch for Microsoft and Sun to concoct some tough frontal attacks on IBM, their avowed common enemy.

Welcome to the Evil Empire, Sun. Hope you end better than Darth Maul, sacrificed for that scum Attakin, but I wouldn't count on it.

What they need to be doing is continually looking for whole new product categories. Continual innovation. Why did they stop after Java? This plan is just an excuse to sit and yell at IBM and Linux for being so damned innovative and providing so well for their customers.

Posted by Jon Kay at September 22, 2004 11:36 PM