Why Would Sun, Microsoft and IBM Debate the Future of Windows, Linux, Solaris?


Jim Zemlin of the Linux Foundation and representatives of Sun and Microsoft will debate the future of their "respective" operating systems in April 2009. That's Linux, Unix and Windows, respectively, if you need a scorecard after all these years. Jim sort of represents IBM, HP and other systems suppliers tied to Linux because they are the major sponsors of the Linux Foundation, of which he is executive director. Intel, with no dog in the hunt, is sponsor of the actual event called the Linux Foundation Collaboration Summit.


It's good theater and a way to build interest in the event. But does it matter?


Yes and no. In terms of managing traditional hardware resources -- the primary job of an operating system -- the three product sets are approaching parity. It's kind of like GM, Toyota and Benz (you decide which represents Linux, Windows and Unix) debating the future of their respective internal combustion engines. Except that they left the major transaction processing TP engine maker, IBM, and the major visual engine maker, Apple, out of the debate. So, no, it really does not matter from that perspective.


But new types of automotive power plants and vehicles are on the horizon. Electric. Hydrogen. Similarly, new kinds of information technology hardware resources are on the horizon. (I'm actually not sure if hydrogen requires a different type of engine but hopefully I'm making the point: operating systems must change to adapt to new hardware.)


Plus, there are all kinds of other changes in hardware resource management, such as surface computing, closer than the horizon. Plus virtualization (like hybrid cars) is already right in front of us.


  • Linux/Unix is probably ahead of Microsoft in taking advantage of server virtualization. Of course, the IBM TP operating system has been there for 30 years (and don't mix up server virtualization with memory, storage, desktop and application virtualization).
  • Conversely, Windows is ahead of Linux/Unix (really one and the same, of course, sort of making the debate lopsided) in enabling surface computing.


As explained in <strong>this recent blog po</strong>s<strong>t</strong>, we have all these different operating environments (still a much better situation than 25 years ago) because of the differences in hardware and what we expect from IT. One size still does not fit all. So yes, the debate should be about what IT users like you want from your systems/storage/networking resources, where you want to get to and how fast and at what cost. Then the guys that build the engines can debate number of cylinders, fuel injection efficiency schemes and the like.