Microsoft made a series of announcements last week regarding the future of the Windows operating system. After hedging the dates on the release of the next-generation versions of Windows for the past few years, it laid out a roadmap and time frame for the product line that extends almost into the next decade.
On the surface, this sounds like a good thing. One thing that drives the ISV and user communities absolutely crazy is the period of time between when Microsoft is “supposed” to be releasing a new operating system and when the OS actually ships.
This is especially true with the server operating system. Waiting for the next version has delayed many a corporate deployment schedule or forced an unexpected, unwanted upgrade process on the core servers in the network enterprise.
Micorsofts plans to focus its energies on three core initiatives will have significant impact on the way corporate networks make use of the Windows Server platform.
While the three initiatives—.NET, Dynamic Systems and Trustworthy Computing—are all important, its the latter two that will have the most impact on the use and deployment of Windows Servers.
Microsoft has talked a lot about its Trustworthy Computing initiative, and Ill leave that for the Security Center folks to analyze, except to say that a secure network operating system is a key component in any corporate enterprise.
The Dynamic Systems Initiative, however, will have direct impact on the way Windows Server is used in your enterprise.
With the release of Windows Server 2003, Microsoft started a product positioning that had specific versions of Server 2003 for tasks such as file, print and Web services.
With DSI, the process goes much further—making possible a much more granular role for a server. The more granular server, such as a DNS server, would run only the code necessary to support that specific role, making for a much leaner, higher performance dedicated server.
The first of the roadmap products that Windows Server administrators should be keeping an eye out for is Service Pack 1 (SP1) for Windows Server 2003, which will include new security tools and many of the security technologies Microsoft is already showing in Windows XP Service Pack 2 release-candidate 1 (SP2 RC1). Microsoft is also expecting a 10 percent performance improvement over the base version of Server 2003.
Going to 64 Bits
Administrators of large-scale database systems will want to keep an eye out for Windows 2003 Server for 64-Bit Extended Systems. This version of Windows Server will use the 64-bit Advanced Micro Devices Inc.s and Intel Corp.s x86 architecture CPUs (not Itanium) and will be based on the SP1 version of Server 2003.
Microsoft has made claims of performance improvements over 32-bit Server 2003 ranging from 17 percent on large databases to more than 100 percent for Active Directory throughput, which, if true in real-world use, will make this 64-bit version a must-have for large enterprises that rely heavily on Active Directory.
These are the short-term changes that Windows network administrators will need to concern themselves with before they even need to start worrying about the Longhorn announcements—which wont become product any sooner than 2007, with the server version shipping nine months after the client version.
And frankly, worrying about Longhorn now is just a waste of energy for anyone other than developers. Given the way technology changes, it would be very surprising if the version of Longhorn that ships in 2007 will bear more than a passing resemblance to the preview versions available today.
One of the things I liked most about the announcements was the “we will ship it when its ready and not a moment sooner” attitude that seemed to pervade the Windows roadmap.
Its always made more sense for the products to ship when they were ready and complete rather than attempting to adhere to some arbitrary release schedule, regardless of how much more marketing-friendly product release date announcements can be.
Of course, then I read Peter Gallis piece in which Jim Allchin talks about syncing up the release dates of Longhorn client and server—a process that failed miserably when attempted with Windows XP and Server 2003.
It would seem that there is still a vocal group at Microsoft that likes the idea of syncing the releases and doesnt shiver every time the phrase “cutting features to make a deadline” is uttered.
Frankly, I have little concern about getting my client and server releases of the operating system at the same time. A staggered release gives me, as an administrator, that much more time to plan deployments and get the architectural changes in place to take advantage of new operating-system features.
Ship it when its ready. Anything else is just a disservice to the buyer.