One of the things I love about this job is the opportunity to discuss industry issues with really smart people, whether it's customers, vendors, or my colleagues in Data Center Strategies at the Burton Group. Recently, Drue Reeves and I have been discussing the imminent arrival of Windows 2008 Server and what it means for the future of Windows Server. We can't agree at the moment, so we're going to have a little Presidential Debate here (though I'm not sure Drue is entirely happy with the role of Hillary ) and hopefully we'll get some feedback from you!
Anyway, the heart of our disagreement is whether W2K08 represents the end of the road for big new operating system releases from Microsoft, or simply business as usual with W2K11 (or 12 or whatever it turns out to be) rolling along with the inevitability of the sunrise. So, I'm going to tell you why I think it is business as usual.
- First on my list of reasons for W2K11 is that the underlying hardware platform is changing rapidly, probably more rapidly than it has for many years. The issue here is that when you design a new operating system, you inevitably optimize its feature set for what you expect the typical hardware to support when the OS is in its "prime". For example, Windows 2000 typically ran on systems with 2-4 processors, optimizing the design for more than 4 processors would have been a wasted effort. My guess is that W2K08 is probably optimized for 16 processors (4-socket system with four cores/socket). But by 2014 (when W2K11 would be hitting its stride), that four socket system will probably look like at least a 128-processor system to the OS (4 CPUs, 16 cores/CPU, 2 simultaneous threads/core). That level of parallelism in the hardware will need to be exploited effectively which is almost certainly going to result in some pretty big architectural changes for the OS. That’s just one of the changes happening to the underlying hardware, and it’s safe to assume that a x86-server in 2014 will be substantially different to anything on the market today.
- Second on my list is Microsoft's stated intent to make Windows Server more modular, they've started down that road in W2K08, but they are a long way from where they want to be. Getting to a fully modular operating system will almost certainly require at least one more major release if not more.
- Third on my list is the fact that Windows Server is a huge cash cow for Microsoft, and I can't see them putting it out to pasture any time soon. Yes, initial uptake of W2K08 will be slow, but then it's been slow for W2K, W2K03, and W2K03R2 before and that's never stopped it from happening in the end. My guess is that within two years, W2K08 will be the default choice for all new Windows Server installs, and W2K03 will start to fade away. One big reason for uptake of W2K08 is that 64-bit computing on the x86-platform is about to come of age, and W2K08 is simply a better platform than W2K03 on 64-bit hardware.
- Fourth on my list is that enterprise IT requirements change over time and the operating system has to adapt. If it doesn't adapt, then IT organizations will move on to something that does the job more efficiently. I don't see this dynamic changing, so there will always be "work to be done" for operating system vendors as they respond to changing IT needs and to their competitors.
So I don't think there is anything that will stop the OS upgrade train, and enterprise IT organizations will continue to buy a ticket to ride.
Anyway, enough for now, I'll let Drue tell you why W2K08 is "The Last of the Mohicans"...
Posted by: Nik Simpson