Software, Interrupted

The fact that SGI was acquired by Rackable Systems for the sad sum of $25 million was big news on Wednesday only because most people had forgotten that the company, formerly better known as Silicon Graphics, still existed.

So what killed SGI? In addition to the rise of Nvidia and makers of other graphics chips that ran on cheaper hardware, it was bad choices:

• Continuing to stick with its own chips, operating system, and hardware while the rest of the world moved to commodity x86 boxes.
• The adoption of Intel's Itanium chip, which remains a depressing joke of a product.
• A myopic view that the core market in high-end servers wasn't being decimated by the rise of Solaris and Linux

When I first started working at Bell Labs in 1995, most of the servers we used were from SGI. People liked the stability of Irix and the fact that the machines looked cool (ask anyone what they remember about SGI, and they will tell you the gear was cool.)

In 1996 or so, I started seeing mass quantities of Sun Microsystems' machines coming into Lucent, and by 1998, much of the company's operations ran on Solaris machines. Sun commoditized SGI much in the same way that cheap x86 Linux boxes have subsequently commoditized Sun.

Can Sun avoid the SGI trap? Maybe, but probably only if the company gets broken up into different business units that separate out the OS, software and hardware.

Sun has some good stuff. Solaris is great, the hardware is fine, if a bit expensive, and the company has good cash flow and products in better market segments than SGI did.

But Sun also has a lot of junk. There is a vast array of Sun software that costs a lot to maintain but doesn't deliver much revenue. This is arguably the area in which Sun's strategy has been so off the mark. With the exception of MySQL, there aren't many Sun software products that generate significant revenue. If software is crappy, and people don't want to use it, then it doesn't matter if it's open source or proprietary.

It sounds like a simple argument to suggest that x86 Linux killed SGI and is killing Sun, but the truth of the matter is that both companies could have made things significantly better for themselves by embracing Linux early on, instead of fighting the tide and waiting until their market position had been vanquished.

The other interesting aspect here is that Microsoft has proven to be the smartest of all, having realized that the decoupling of the operating system, hardware, and applications would eventually give it the ability to dominate multiple markets by owning the user platform.