The Upgrade Dilemma

Microsoft needs to make upgrades easier

The question of whether to standardize on Microsoft products has always been an interesting one for companies. Historically, Microsoft has combined low prices with ease-of-use advancements to make its offerings more compelling than the competition's. But along with steadily increasing its market power in recent years, Microsoft has slowly tightened the licensing vise, turning what were once obvious product upgrade decisions into dilemmas that require careful study.

Win2K and NT 4.0
Because Windows 2000's acceptance is increasing, Microsoft will end Windows NT Server 4.0 support in January 2003; after this time, free hotfixes won't be available. In January 2004, NT Server 4.0 hotfixes won't be available at any price.

Unlike NT 4.0, Win2K supports Microsoft Windows Update and Microsoft Office Auto Update, and Microsoft will continue to upgrade the OS with service packs for some time. But Win2K is such a huge technological jump over NT 4.0 that upgrading to it is often difficult, especially for users migrating complex NT 4.0 domains to Active Directory (AD). Indeed, some enterprises have stuck with NT 4.0 because of the scope of a Win2K upgrade. Many others have simply rolled out Win2K on new systems only and retired—rather than upgraded—NT 4.0 boxes as they did so. Either case can be a support nightmare.

License 6.0 and .NET Server
This year's release of Windows .NET Server throws yet another factor into the upgrade dilemma. .NET Server will be accompanied by License 6.0, which Microsoft briefly launched in late 2001, then pulled because of customer complaints.

The problem with License 6.0 is that it essentially requires corporations to upgrade to new Windows versions as they're released—every 2 years in Microsoft's plan. But enterprisewide deployment of a new OS requires substantial resources, and many companies operate on a 3- to 5-year software-upgrade cycle, not a 2-year cycle. Microsoft seems to think that most users simply need to get used to the new policy—not that it's inherently flawed. The company has extended the current licensing program to July 31, but License 6.0 won't likely go away. (For more information about License 6.0, see Kathy Ivens, "License 6.0: The New Deal," http://www.winnetmag.com, InstantDoc ID 24032.)

The Future: Subscriptions
Perhaps software subscriptions, which would institute the type of regular income Microsoft so desperately desires, will eventually solve the upgrade dilemma. A subscription model could nudge Microsoft toward more frequent, smaller upgrades because its revenue would no longer be tied to large product upgrades. Consider the 4 years between the introductions of NT 4.0 and Win2K: If a subscription model had been in place during that time period, Microsoft could have rolled out Win2K features such as IntelliMirror, AD, Microsoft IIS 5.0, Plug and Play (PnP) support, and power management as it completed them, rather than packaging them all in a new OS version. Users might have picked only those technologies they needed, and Microsoft could have continued to actively support just one platform.

Making a big platform leap from time to time is probably necessary, and the move to the .NET distributed computing model is inevitable. However, I believe that Microsoft should proceed in small upgrades over the next few years, not through a major product upgrade that changes everything and renders obsolete all the company's previous enterprise products. This type of rollout would benefit both Microsoft and its customers and wouldn't undermine our belief that the Windows platform is still the place to be.