I predict that Linux will eventually be at the foundation of nearly every enterprise system and that the whole issue of which server operating system to choose will then disappear into ambient background noise. It's not often that I make predictions about predictions, but because the above prophecy is so bold, I'll make an exception: I predict that this will turn out to be one of the easiest predictions I've ever made.
Here's why this prediction will come true: While there are many technical challenges ahead for Linux, there is one overriding factor that makes Linux virtually unassailable by all closed-source competition: Your investment in Linux is protected by the best software warranty on earth, the GNU General Public License.
Here's when the turning point will occur. One IT decision-maker someone somewhere in Fredonia, Iowa, I'm guessing will have the epiphany that he has been asking the wrong question about how to protect his company's software investments. You know the question: "Will this vendor be around to support my installation 10 years from now?" That person will represent the tipping point of all IT decision-makers, and then the revolution will come.
The revelation that awaits these folks is that this question is absolutely meaningless when it comes to closed-source software. All closed-source software is a time bomb. Anyone who doesn't hear the ticking is in denial. There are only two types of closed-source software vendors: those whose software becomes obsolete and impossible to support, and those whose software becomes obsolete and impossible to support but that have a better sales strategy. The former group consists of vendors that will go out of business. The latter group includes vendors that will be in business 10 years from now because they've convinced enough customers that the last software version they were using is now obsolete and impossible to support, so they should invest in the new version.
This makes for a delicious paradox if ever there was one. The longevity of a company may be proportional to its market share, but it's inversely proportional to how soon your software will become obsolete. Microsoft Corp.'s attempt to move to a subscription-based software model is a perfect example. The company is basically saying, "We're going to make your software obsolete so often that you might as well just pay us for 10 years' worth of software upgrades upfront and get it over with." Any company with less than 90 percent market share would be laughed off the face of the earth for suggesting that. And that may be just the reaction Microsoft gets. But it couldn't even float the idea if it didn't already own the desktop.
Here's when the prediction will come true. Once that Iowa guy comes on board, IT decision-makers will begin to ask the right question: "Do I get the source code and the right to modify and redistribute it at will?"
Linux and most of the code that makes Linux useful are licensed under the GNU General Public License, which means you're guaranteed to have access to the source code no matter what happens to the company that produced your flavor of Linux. This means that there's no such thing as a doomsday scenario for Linux.
Assume, for the sake of argument, that all Linux distributors went out of business and that even the all-volunteer projects like Debian ceased. So, what would happen when your next hardware upgrade broke the Linux kernel? Nothing much. You'd have the source code and every tool you'd need to fix the problem. All you would need is a programmer. If that sounds unpleasant, imagine what would happen if Microsoft folded and your next hardware upgrade broke Windows. Without the source code, you're basically out of luck. Of course, this is purely hypothetical, because you'd already be paying half your staff to deal with the consequences of the latest migration to the next version of Windows.
My prediction is reliable for other reasons. For example, unlike Microsoft Certified Systems Engineer graduates, Linux experts have complete control over their own value to a company. Nobody can decertify a Linux expert, because nobody can claim exclusive ownership of the intellectual property they learn. Every reason I can cite to support this prediction points back to the same issue having access to the source code.
So if you happen to be that IT decision-maker in Fredonia, Iowa, please let me know when you come around, so I can pass it on that the revolution has passed the turning point.