Windows 2000: And the Survey Says...

Analysts have been lobbing grenades at Windows 2000 (Win2K) for the past year. At the Windows NT in the Enterprise: Is IT Up to the Challenge? conference in May, GartnerGroup summed up the major points of Win2K criticism. GartnerGroup based its arguments on the analysis of Win2K's total cost of ownership (TCO) and a calculation of the Return on Investment (ROI) that an enterprise can expect for Win2K. At the heart of the GartnerGroup analysis were the imposing administrative requirements of deploying Win2K. Some of GartnerGroup's projections are

GartnerGroup recommended that "organizations shouldn't deploy Win2K at the desktop for a midsized NOS until Microsoft has delivered the first major proven service pack (or 6 to 9 months after the general availability of Win2K)." Although GartnerGroup's analysis was in the early Win2K beta time frame (when Microsoft still called the product Windows NT 5.0), GartnerGroup's conclusions likely haven't changed much. GartnerGroup offers prudent advice and makes the same recommendation that Windows NT Magazine has made in the past year.

Assessing TCO arguments is difficult. To create TCO projections, you build a model (GartnerGroup built the original model and coined the term TCO); plug in the hardware, software, and cost of implementation; then assume a certain system lifetime. Hardware and software costs are generally known, but lifetime and implementation costs are highly variable. GartnerGroup's analysis tool creates a slider that you position based on how your organization uses computers. Then, the tool calculates the TCO.

Hidden from any TCO analysis are the business opportunities created by having (or lost by not having) a particular OS running in your business. Because an OS is enabling technology for your business, if you don't buy the OS, you can't upgrade to a new make-or-break database management system (DBMS), messaging platform, or accounting package. How can we measure such a loss? For this reason, I don't place full trust in TCO arguments regarding OSs.

However, many companies, including Microsoft, take GartnerGroup's TCO calculations seriously. Microsoft's Zero Administration for Windows (ZAW) initiative, which is a suite of technologies ranging from IntelliMirror to remote installation and network booting, is a TCO argument to counter GartnerGroup's claims. Microsoft didn't send a speaker to the conference, but Microsoft did reply to questions after the conference.

Design goals, TCO, and ROI aside, Windows NT Magazine wondered what businesses thought about Win2K's prospects upon the OS's release: How do businesses plan to deploy Win2K and to what extent? Windows NT Magazine selected nearly 8000 systems administrators whose demographics identified them as decision makers for Win2K deployments in organizations of various sizes. Then, Windows NT Magazine turned that list over to World Research to survey.

The final sample group included about 1400 people and contained 88.2 percent IT professionals, 26.3 percent business executives or managers, and 19.6 percent software developers. (The total is more than 100 percent because respondents could select more than one designation.) The business makeup that the respondents represented was 42.1 percent small businesses (i.e., 1 to 249 desktops), 33.1 percent midsized businesses (i.e., 250 to 4999 desktops), and 24.8 percent enterprise businesses (i.e., 5000 desktops or more).

Peter Auditore, vice president of the Syndicated Services Business Unit of World Research, ran the Win2K deployment study. Auditore drew the following conclusion from the study: "People are going to rapidly upgrade to Win2K, and by the end of 2000, the bulk of deployment will have happened." This study group expected to deploy Win2K after 8 months of assessment. World Research found that 44 percent of the group began assessment with Win2K beta 3 and 60 percent expected to complete planning and budgeting by first quarter 2000.

If the pattern that the survey suggests occurs, Auditore said, "Microsoft could make $10 billion in 2 years for just upgrades to Win2K Professional, and maybe $10 billion more for upgrades to Win2K Server." The survey results suggest that by January 1, 2001, 90 percent of desktops will run some form of Windows and 40 percent will run Win2K Professional (Win2K Pro). Therefore, the penetration of Windows on the desktop won't change much during the upgrade period. The same organizations will have 81 percent of their servers running Win2K Server or NT 4.0. This survey largely represents Windows NT Magazine's readership, which generally supports Windows OSs. World Research measured little if any crossover to competitive OSs among this group.

Microsoft emphasizes three reasons that you'll want to upgrade to Win2K:

Microsoft also mentions other benefits that Win2K offers, such as Kerberos, public key infrastructure (PKI) certificate security, and improved performance.

The World Research survey tested how well Microsoft's message about Win2K's benefits was resonating with this particular audience. The survey group rated reliability as the most important feature and primary concern. Features of low interest were Active Directory (AD) and TCO. When asked to rank scalability, only participants in large enterprises ranked scalability as even moderately important.

When asked to describe how they will use Win2K in their companies, systems administrators and IT managers listed usage in the following order: file and print, databases, Web servers, messaging servers, and data warehousing. This pattern suggests that this group is looking at Win2K as an upgrade deployment.

The survey also provided some surprises. A remarkably low number of enterprises are designing around AD. Auditore said, "I don't think Microsoft has marketed this OS correctly. Microsoft hasn't fully described the value of AD."

The survey also measured a very low expected deployment of Win2K Server Terminal Services. Auditore felt that the plummeting cost of hardware for desktop OSs has lowered Terminal Services' TCO benefit and the potential effect of the thin-client/server model. To that, I add that the early pricing policy of double licenses for Citrix MetaFrame running on NT 4.0, Terminal Server Edition probably wasn't helpful.

Because the majority of the respondents expected to complete their Win2K implementation plans before the release of Win2K Datacenter Server (Datacenter), this survey clearly hadn't captured that group of decision makers who will likely evaluate and decide on Win2K's enterprise prospects. The survey measured the NT installed base's desire to upgrade, and we'll have to wait to learn what happens beyond this straightforward upgrade activity.

According to the survey, Win2K is likely to be one of the most successful OS introductions in history, with a very rapid deployment in Windows-based shops. The business managers and systems administrators that World Research measured said they plan to bite the bullet, train their staff, upgrade their equipment, and take the financial hit to deploy Win2K. Well, that's what the survey says. One of Windows NT Magazine's technical editors who disagreed with these projections commented, "I think Win2K will be the most ignored OS Microsoft has ever produced. Win2K might be the product that kills Microsoft." My opinion is that Win2K is likely to be a financial success for Microsoft but probably not a technological success. Microsoft is imposing severe hardware requirements on Win2K adoption. This factor means that a soft underbelly exists that other OSs and technologies can exploit. With Win2K, the door is wide open for administrators to install Linux and other server appliances on legacy PC servers in the enterprise, and both Linux and other server applications are going to get play in the market as administrators deploy Win2K.