Sun now sells Windows servers.

Whisky. Tango. Foxtrot.

Let’s lay this out: In the mid-to-late 90s, Sun was the go-to vendor for high-end server hardware, PLUS they had the best commercial Unix by a damn sight. Solaris — poorly named though it was — had it all over the similar offerings from HP (HP/UX), IBM (AIX), and others now lost to history. Add to this their visionary — at the time — pursuit of a cross-platform programming language more or less “native” to the Internet (Java), and you got an image of a company really going places.

Then a couple things happened.

First, Linux happened when a Finnish kid managed to assemble a workable Unix clone from the Gnu project and a new kernel. It started out ok, but improved VERY rapidly; in the 16 years since its first released, it’s become one of the most popular Unix choices even in conservative enterprise situations; of the proliferation of commercial Unixes that Solaris competed against in the 90s, only a few remain, and most of them are weak indeed. Linux gives the user freedoms you don’t get with proprietary, closed choices, and can now boast of a developer ecosystem that pretty much dwarfs that of any closed-source server OS. Even Solaris is now Open Source (or has an OS version), since that’s the only way to compete with Linux.

Second, the commodity Intel box got more and more and more powerful, meaning folks didn’t NEED Sun hardware to do things like run mail and web servers, or even databases, for most applications. Serious hardware just isn’t indicated in 99.9% of situations — and even when it is, Intel architecture boxes have grown up into high-availability tools that cost a fraction of what Sun’s Sparc hardware demanded.

The combination of a free operating system better than Solaris and cheap hardware getting more cost-effective than Sparc has pretty much cut off Sun’s oxygen supply, to steal a phrase.

It’s amusing that Sun’s lasting legacy is Java, but also predictable. Sun realized that for Java to succeed, it needed to be essentially free. Sun makes almost no money from Java, and can’t — if they charged for it, they’d kill it. The era of pay-to-play development environments is just plain over. Java’s still huge, and will likely stay that way for a while, but it’s impossible for Sun to monetize that success. (Even so, the rise of dynamic languages and frameworks will likely mean Java’s time as king of the hill is limited.)