What could Microsoft learn from Linux? Jason Perlow takes a look at Windows Server 2008, and says that Microsoft should consider taking a page from its competition, the Linux distro vendors, and consider shortening the product release cycle to allow for more incremental, short term upgrades.
By the time this column sees the light of day, Windows Server 2008 will be RTM’ed — Released to Manufacturing. Oh joy.
Now, don’t get me wrong here, I am one of the few Linux die-hards that actually LIKES Windows. I’m sorry, I can’t help myself, it’s a disease. I’ve said it before, Windows is in my blood — I spent a lot of my early career doing Windows support and systems engineering at Fortune 500 companies, and I still use Windows as my primary productivity platform. I also feel that Windows Server is a perfectly good commodity computing operating system, if best practices for patching and keeping the systems maintained are applied.
Still, I don’t see a lot of big companies chomping at the bit to go upgrading to Windows Server 2008. Certainly, there are several things to like about the new release, such as improved security and disk encryption inherited from Windows Vista, a streamlined management interface, enhancements to clustering, built-in virtualization, integrated IIS 7.0, componentized installation, an optional” headless” operation mode and role-based configuration.
However, with the notable exception of the aforementioned Hyper-V virtualization (which, by the way, will eventually be available as a de-bundled separate product for the low price of about $28 and will be compatible with Xen-enabled Linux VM’s — very cool, although neither VMWare nor Citrix should be shuddering yet) all of these things are very incremental enhancements, and do not result in very compelling reasons for a quick mass adoption or migration to the latest version of the Server OS.
Frankly, and I hate to say it, there are a lot fewer reasons for wanting to go to Windows Server 2008 from Windows Server 2003 than there are to migrate to Windows Vista from Windows XP, which pretty much every pundit covering this industry would agree has been a less-than-stellar adoption.
Exchange 2007 doesn’t require Windows Server 2008 and neither does SQL Server 2008, which represents two of the three primary application server roles for the OS. IIS 7 does require Windows Server 2008, but it’s questionable if IIS 7 could be considered Windows Server 2008′s” killer app”. As with Vista, which also has no” Killer app” as of yet, it could take a long time before someone creates that super-compelling. NET application that absolutely requires Longhorn and IIS 7.
As we all know about server OS adoptions in large corporations, migrations and upgrades are measured in five to seven year cycles, and the current Windows Server migration hasn’t even peaked yet, as there are still many Windows 2000 Server boxes still in the field that are only now being displaced by Server 2003 — and this is being driven by the fact that” extended” Windows 2000 support will end in June of 2010. When Microsoft ends Windows 2003″ mainstream” support — which will not happen until at least sometime in 2010 — and with Extended Support dying in 2013, Server 2008 may eventually start to look attractive.
But three to five years from now is an awful long time away in the computer software industry, and you can bet that Linux — which is currently displacing a big chunk of the Midrange and Enterprise space that UNIX once reigned supreme — will also be chewing up more and more of that commodity server computing space as the management tools, compatibility and Windows interoperability of Linux and other Open Source software vastly improves. You can thank the most recent developments in European Commission-driven SAMBA/Microsoft agreement for much of that improvement.
Sure, there are controls and restrictions on how the SAMBA organization can receive that information, but for all practical purposes, many of the key barriers to compatibility and interoperability between Windows and Linux at the server networking level will be coming down. And that will translate into some smart CIOs thinking about moving some of those roles to Linux instead of Windows Server 2008.
Perhaps Microsoft should consider taking a page from its competition, the Linux distro vendors, and consider shortening the product release cycle to allow for more incremental, short term upgrades as we see with RHEL and SLES, and certainly as we see with much more frequency in community distros such as Fedora, openSUSE, and Ubuntu. With Windows now becoming ostensibly more” modularized” through services such as Microsoft Update, multi-year waits between major release periods in order to add newer or improved functionality should become a thing of the past.
It’s going to be an interesting couple of years for everyone involved, to be sure. Whether Microsoft will able to maintain its dominance in the face of these challenges, or if Linux vendors and Open Source Software projects are able to take advantage of this opportunity to displace Windows market share in the commodity server OS space remains to be seen.