ANALYSIS: Windows 2000 in 2000?

By Chris Gloede

What's the deal with Windows 2000? It looks as though Beta 3 is on time and the product will ship by the end of 1999. The question is, at what expense?

Upgrading an operating system or data base engine is no easy task. There are many issues to evaluate before installing. Major issues are the quality of the release, the stability, the documentation, the portability, and the upgradeability. Let there be no mistake, this is a lot of stuff. Some of which Microsoft has not been known for in a positive light.

Microsoft has been diligently trying to add more and more value to Windows 2000, value like Active Directory. As a result Windows 2000 will involve millions of lines of new code. Millions of lines that need to work well before the product is released. A new operating system that needs to function on a variety of platforms and install easily. All this while the market is anxiously awaiting the long delayed release.

While attempting to add all of these new features and functions, Microsoft has stated that the product will not be released "until it is ready", indicating that it will be a stable release. This means that the Redmondians are spending a great deal of time, money, and effort working on and correcting bugs (opportunities to software developers). What this leaves at this point is an uncertain upgrade path.

There are many questions to be answered before making the jump to Windows 2000. These assume, of course, that the application itself will be stable and worthy of the effort. Given Microsoft's new initiatives in quality of product, pre-release, I believe that this will be the case. Two questions that spring to mind immediately are: Will Windows 2000 run on all of the hardware that Windows NT 4.0 with Service Pack 4 runs on today? Will all of the applications currently supported by NT 4.0 w/SP4 be supported by Win2K? Where the answers to these questions are not an iron clad "yes", will there be sufficient documentation before Installation about what will and what will not work? Will there be a utility that I can load that will examine hardware and applications and produce some list of potential problem areas or shall we just "wing it"?

Remember when IBM released the CISC to RISC upgrade. The operating system V3 was released and the upgrade path was extremely well documented. There was a utility to run that would examine the applications, source and object on your existing system and warn you of impending problems. The actual upgrade process was offered in several variants like upgrading on the same machine or doing a side by side upgrade. The process was long and tedious and required a great attention to detail so that nothing was skipped, but it did work and work well. Most AS/400 operating system upgrades are significantly simpler than the CISC to RISC processor and V2 to V3 upgrade. This one was a major change in so many areas that the risk could have been extreme. IBM handled it about as well as you could.

Will Microsoft do the same with the NT 4.0 to Windows 2000 path? Will there be a recovery other than zapping disk and starting over? I seriously doubt it. Microsoft has not been known for paths forward or backward. I cite Visual C++ 5 versus 6 as an example. It is a safe bet that a fair amount of hardware will be lost along the way and that there will be no easy path to recovery. I tried to reach someone in Redmond to talk to about this but had no luck.

Before upgrading all of your servers, I strongly suggest that you have solid, tested backups and a fair amount of time. I also suggest testing on some non-critical machines. One more thing, a little prayer certainly couldn't hurt.

A veteran of the IBM midrange arena since 1983, Chris Gloede is executive VP for Business Solutions Group in Wayne, Pa.