Recently, Windows NT Magazine partnered with World Research (http://www.survey.com) to survey more than 1400 IT professionals about their plans to adopt Windows 2000 (Win2K). The survey results indicate that almost half of you will have done your technical evaluation, planned, and budgeted for your migration to Win2K by the end of 1999. (For more information about this survey, see my editorial in our UPDATE email newsletter at http://www.winntmag.com/ update/archive/ view.cfm.) This survey shows that Windows NT administrators are testing Win2K, and they have clear priorities in their concerns about implementing it.
We asked survey participants to rank their concerns about Win2K. The top concerns in order of priority are reliability, performance, security, ease of administration, scalability, Active Directory (AD), and reducing total cost of ownership (TCO). Let me address the four of these concerns that I find most compelling at this point: reliability, scalability, AD, and TCO.
Almost 90 percent of the survey respondents listed reliability as their top concern. Recently, NT's reputation as a reliable network OS (NOS) has taken a beating. In my June editorial, "Chasing 9s," I claimed that a component of NT's unreliability was third-party device drivers. That statement prompted a flood of responses from readers recounting their personal experience with NT's reliability problems and giving Microsoft 100 percent of the blame. NT customers are angry, and several have decided to move some of their servers to Linux in search of reliability.
Only 36 percent of our survey's participants listed Win2K scalability as a top concern. For many IT professionals, NT is scalable enough. Sure, NT has room for improvement. But for more than 80 percent of respondents, NT's scalability is good enough. Most people assume that Win2K will be more scalable than NT—if nothing else, as a result of hardware improvements, which have been following Moore's Law. (For example, in the past few months, tests of new 8-way servers with SAP R/3 and Microsoft SQL Server 7.0 have demonstrated dramatic increases in NT Server 4.0's scalability.)
AD will be a significant aspect of Win2K, and because of AD, one impetus for migrating to Win2K will be the next version of Microsoft Exchange Server, code-named Platinum. Platinum replaces all of Exchange Server's proprietary directories with Win2K's AD, so that Platinum requires a full implementation of Win2K to run. (In his May article, "Discovering Platinum," Tony Redmond explains Exchange Server's new features and their significance.)
Although reducing TCO with Zero Administration Windows (ZAW, now called Change and Configuration Management—CCM) and IntelliMirror was one of Microsoft's primary focuses for Win2K, the survey participants listed TCO last among their priorities. Why the low priority? Perhaps many IT professionals have worked around NT's shortcomings instead of waiting for Win2K to fix them. Perhaps the term TCO has been used (and abused) so much that it has lost its sizzle. Or, perhaps the biggest factor in TCO comes back down to reliability: Downtime is expensive, and if the system doesn't stay up, who cares how easy or cheap it is to administer?
Reliability: The Prime Directive
If Win2K proves to be more reliable than NT 4.0, people will migrate for that reason alone. If Win2K proves to be unreliable, watch for erosion in NT's market share. As I've written in the past, reliability was the foundation of NT's success. Back in my March 1997 editorial, "The Prime Directive," I predicted that if NT were to lose its reputation for stability, NT's momentum would stall, and this prediction has come true. In part, Linux's 17 percent growth has resulted from lack of faith in NT's stability.
Microsoft is under pressure to release Win2K soon, but, as I said in that 1997 editorial, I urge Microsoft not to release Win2K before its time! Reliability is still the prime directive, and it's worth the wait.
What's important to IT professionals in Win2K?
4. Ease of administration
7. TCO reduction