Pivotal Time for Servers

Now is the time for NT administrators to balance the costs, freedoms and risks.

Corporate IT is at a network operating system crossroads. Its decision time for Windows users, who need to make hard choices about which road will lead to where they want to go. With the shipment of Windows Server 2003 last week, Microsofts direction in its server line is clear: The Windows 2000 Server base and feature set, which have been with us since February 2000, will be the core of Microsofts value proposition on the server for the next two or three years as well.

Windows Server 2003 doesnt rearrange the competitive landscape. Its a refresh release to keep the server platform in line with hardware progress and to entrench .Net Framework while Microsoft works on more fundamental advances for its "Longhorn" release. In fact, many of the development and security features in the product can be deployed on Windows 2000 Server today through free downloads Microsoft has posted over the past few years.

With Microsofts path laid out, now is the time for Windows NT 4.0 customers to choose whether they want to commit to the Windows 2000 path or if Unix or Linux alternatives are better matches for their technology, staffing and economics. Those who rely on the Windows 2000 core technologies—most notably Active Directory and group policies—and are using emerging Microsoft language technologies such as .Net Framework have made a choice in favor of the seamless integration and powerful development tools that an end-to-end Microsoft approach provides. In this process, Windows Server 2003 is a logical next step.

However, a real decision faces the mass of Windows NT Server administrators, who keep clawing their way back from the brink of product abandonment one Microsoft support extension at a time. Microsofts messages have been clear. Its initial announcement that no NT security fixes would be offered after the end of this year was met with a hail of protest, and the date was later extended by one year.

Microsoft makes support decisions based on its business needs, but because it is the only company able to fix the closed NT code base, when it pulls the plug, NT is done. Microsoft could, and should, use the virtual machine technology it recently acquired from Connectix to allow NT applications to run on Windows Server 2003; using virtualization technology to do the same thing on a non-Windows platform is just as possible. Given the changes Windows 2000 introduced, a transition from Windows NT to Windows 2000 is in some cases comparable to a migration from Windows NT to Linux or Solaris. Users using Windows file and print features, or running packaged applications from non-Microsoft companies, most of which are OS-agnostic, will find a cross-OS migration quite feasible.

Now is the time for NT administrators to balance the costs, freedoms and risks in their server OS decisions. As Yogi Berra put it, "When you come to a fork in the road, take it." The time to decide is upon us.

For more on Windows Server 2003, see our special section.