As the clock runs down on the final 12 months of Microsoft Corp.'s support for Windows NT 4.0, many IT professionals must make a critical decision: Should they migrate systems to Microsoft Windows 2000 or to Microsoft Windows XP and Microsoft .Net Server? Each choice has benefits and drawbacks.
Windows 2000 is a mature platform that offers proven functionality and low cost of ownership, stability and ubiquity, in terms of third-party support and application compatibility. On the other hand, .Net Server hasn't yet shipped but offers numerous enhancements such as:
- Improved support for Web services (integration with UDDI, GXA and Microsoft .Net Passport).
- Better scalability and replication in Active Directory.
- Domain name server improvements.
- Transitive trusts between Active Directory forests.
- Stronger "out-of-the-box" security.
Conventional wisdom would have it that Microsoft's .Net platform is simply too new, so organizations should deploy Windows 2000 instead. In reality, however, the decision is more complex. While Windows 2000 was a massive upgrade from Windows NT 4.0, .Net Server is more of an incremental change from Windows 2000, and the path from Windows 2000 to .Net Server is relatively simple. In addition, Windows 2000 Server is nearing the midpoint of its lifetime, in Microsoft's five- to six-year life cycle for most of its products. With mainstream support for Windows 2000 scheduled to end in early 2005, do CIOs who will finish deployments in late 2003 want to conduct another upgrade cycle -- from Windows 2000 to .Net Server -- in 2005?
There is no obvious "right" answer to the question, but IT professionals should take into account the entire IT environment (including product life cycles for all hardware and applications), major business functions it supports, the organization's tolerance for risk and current migration plans (or lack thereof). Following are some real scenarios I have heard from companies, and my advice.
Migration to Windows 2000 Already Under Way
"We started planning and testing migration from Windows NT 4.0 to Windows 2000, and we're well into our project. Do we need to drop everything and start over with .Net Server?"
Many companies already invested time and resources in migration, and returns are within reach. In these cases, it's probably not worth scrapping careful planning, and best to complete the process. However, CIOs may want to consider how to "slip-stream" .Net Server into the project, since Windows 2000 Server doesn't present significant differences from .Net Server. They also should take this opportunity to reduce overall maintenance and management costs, through tactics such as server consolidation.
Just Beginning
"For the last two years, we have been consumed with an ERP application rollout and integration projects from a company merger. We know we need to migrate but haven't decided on a direction."
In this scenario, planning should begin immediately; preparation and testing will take several months. Because this will run out the clock on Windows 2000 product longevity, CIOs should note that the .Net platform will mature in parallel with their preparations and provide more use over the (longer) balance of the product's lifetime.
This example also presents an opportunity to achieve other benefits. Like companies that have migration work under way, those that haven't begun should plan for general system enhancements that can be accomplished simultaneously, such as server consolidation and operations management improvements. They should chart a path for systems to support future initiatives, such as Web services, as well.
Scenario: Procrastination
"It ain't broke, so we don't plan to fix it."
Some companies may respond to the phase-out of support for Windows NT 4.0 with something that sounds like pragmatism. However, doing nothing poses risks, too. Industry analysts point out that unsupported software can pose legal liability problems. New hardware drivers for Windows NT 4.0 may not be available. And remaining on the platform may limit a company's ability to deploy new applications and services that support its business.
This isn't to say there aren't sound reasons to stay on Windows NT 4.0. You might have a critical business application that won't run on a later platform. In fact, your answer depends on what applications Windows NT 4.0 supports, your company's likely future initiatives and your organization's tolerance for the risks of using unsupported software.
Christopher Burry is technology infrastructure practice director and a fellow at Avanade, an integrator for Microsoft technology that's a joint venture between Accenture and Microsoft. Readers are invited to send comments or questions to Burry at Christopher.Burry@avanade.com.