What Windows Server 2003 Will Mean for IT

This month’s release of Microsoft’s Windows Server 2003 will be welcome news for some enterprise players, such as in-house application developers and perhaps some Internet service providers. For others, it will be like walking the plank: They may not want to go forward, but they will be unable to go back — or even maintain the status quo.

At least, that is the view of some analysts who have been examining the latest OS from Redmond. While some important features are included in this release, the question, “Should my company upgrade?” is irrelevant for many CIOs. Support for older Windows operating systems is expiring, and in some cases, Microsoft’s volume licensing deals mean companies have already paid for the new software. These two factors will force many a CIO’s hand.

Forced March

Windows NT 4.0, still a widely used platform in many enterprises, will reach its end-of-life at the close of 2003, meaning no more support will be forthcoming from Redmond. Companies that decide to upgrade will find that Win2003 is their only choice. “There’s really no other way to go, save moving to another platform,” Gartner vice president David Smith told the E-Commerce Times. “If you want to continue to use Windows, you won’t really have any other choice after 4.0 is no longer supported.”

In addition, passing on Windows Server 2003 may be like leaving money on the table. Microsoft rolled out its Volume Licensing 6.0 last fall. In order to lock in product prices for two to three years, customers were nudged into committing to minimum numbers of product licenses in exchange for discounts, according to a point system. (If you want to keep track of your brownie points as a Licensing 6.0 customer, software called eOpen from Microsoft supposedly can help track such data, and you can browse through a humongous document of point charts.

As part of Licensing 6.0, Microsoft’s Software Assurance grants free access to upgrades of a product. Therefore, if you are on board with Licensing 6.0, buying into Windows Server 2003 is more or less a matter of picking up what you have already purchased. As Forrester Research principal analyst Frank Gillett put it, “Hey, why not use it?”

Ready or Not?

Gillett is only half-serious. Despite the double inducement of licensing and support, companies still should think carefully about what they are taking on if they choose to transition. Some parts of the migration likely will be good, but some will cause pain.

The most difficult question facing CIOs who are not stuck with NT 4.0 or volume licensing is whether or not the new features and stability improvements of Windows Server 2003 are worthwhile. An IT administrator might want a more reliable operating system, but does he or she want to shoulder huge new projects, such as setting up Active Directory on a company’s production network?

Thorough consideration involves evaluation of Windows Server 2003’s reliability improvements and some new technology. For example, the new product features application partitioning, designed to make crashes in individual Web applications less likely to trash an entire Web server installation. Windows Server 2003 also will be the first server operating system to include support for .NET, Microsoft’s virtual environment for XML-based Web applications.

Misspent Youth

“Windows has gone through a brash youth and is entering its young adult phase,” Aberdeen Group chief research officer Peter Kastner told the E-Commerce Times. In his opinion, “The performance and the stability and the security features of [Windows Server 2003] are far more important than any of the particular new features included.” He said he believes Microsoft delayed the platform’s release to improve stability and added that he is impressed with the TPCC benchmarks he has seen — a measure of the number of database transactions the software can enable.

Gartner’s Smith concurred that the platform’s basic security features are paramount. He rated the partitioning feature for Internet Information Server as an important bulwark against malware and denial-of-service attacks. As for .NET, Smith said those in the know already have it. “You didn’t need to wait for Win2003 to deploy .NET. If you did, you didn’t really understand the software.”

But Aberdeen’s Kastner said that, nonetheless, integration of .NET with Win2003 is important because it makes .NET more reliable. “You have to realize Win2003 is a long-term project to bring underlying stability to a bunch of applications, including .NET, Exchange and Office 11.”

Favorite Features

Some analysts said they have favorite features. According to Gartner’s Smith, improvements to Active Directory, such as the addition of “wizard” templates to streamline policy setting, will be the main feature improvement. And Forrester’s Gillett gave a thumbs-up to the dynamic systems initiative, referred to as ADS. “You will be able to throw up three Web servers to meet extra traffic at peak hours, then reconfigure that machine several hours later for some other purpose,” he told the E-Commerce Times. He believes this feature could save enterprises a significant amount of capital compared with such alternatives as hosting firm Opsware.

At the end of the day, some will want to stick with the devil they know, despite the fact that they have paid in advance for the upgrade, and despite Kastner’s statement that “you can feel confident putting this (Win 2003) in your production network right now.”

Should an enterprise feel trepidation if, for the sake of caution, it decides to hold on to Windows 2000 until that system reaches end-of-life in 2005? Not at all, said Gillett. “You can run NT 4.0 on top of Windows 2000 or 2003, or you can just go without support.” Neither is a great option, but at least users will not be left behind entirely if they decide to wait. “By 2005, there’ll be 4 GHz Xeons, and you’ll be able to run this stuff with fewer chips,” he added.

So, for thrifty companies, as well as those simply unwilling to walk the plank, saying no to the Windows Server 2003 upgrade remains an option. It may not be the easiest choice, but enterprises must evaluate their own needs and then make their own decisions. The new OS will still be there when they are ready.