Having survived efforts to break up its Windows monopoly, Microsoft is now leading the charge to break up the very operating system it defended. With the release of Longhorn in 2007, the company has said it will offer "role-based" versions of Windows in which only the code needed to perform a given function will be included in a particular build of the operating system. It seems that the monolithic, one-size-fits-all approach to Windows may finally be reaching the end of the road.

The idea of a slimmer Windows has been long in coming. It runs counter to the company's marketing urge to add ever more features. But the move toward smaller, simpler Windows versions is driven not by marketing and sales as much as it is by competitive pressure and the demand for greater security, manageability and flexibility. And those trends could well extend beyond simply tailoring an operating system to a few task-specific server functions.

Windows today isn't so much a precision tool as it is a very large Swiss Army Knife, with dozens of features, functions and services extending from its core. But administrators don't want all that in a Web, storage or print server. Leaner, function-specific versions of Windows could present fewer vulnerability points to attackers while making installation, patching and maintenance easier. The strategy could extend to a wide range of server roles and would benefit security-sensitive functions such as call management and Web servers.

Microsoft's approach also represents a calculated response to Linux. The open-source operating system has done well in focused server roles, in part because IT professionals can disassemble the kernel and recombine the pieces required to create a purpose-built version of Linux. In so doing, it gives administrators a lot more flexibility than Windows allows today.

Now, rather than simply selling task-specific editions of Windows, Microsoft may let systems administrators choose which core elements of Windows to include at installation. That's good news for administrators, but the question is how far Microsoft will go in letting customers define and extend those role-based configurations. How much flexibility will IT gain - and how much control is Microsoft willing to give up?

Microsoft might decide to let administrators build optimised Windows versions for a few predetermined roles, such as print or storage server functions, but not allow full control over which core components for other functions are included or excluded. Meanwhile, businesses might want to extend the concept to the desktop, where administrators could create streamlined, role-based deployments for workstations used by retail clerks, warehouse workers or financial analysts.

The role-based approach could also be applied further up the software stack, where the same code-bloat problem infects applications such as Office. The tightly integrated set of desktop productivity applications offers a wide range of features that most people never use. And while some major features can be eliminated at installation, Office remains a large and unwieldy application set for users with fairly simple or narrowly defined roles. Rather than saying "Supersize it" for every application, administrators should have the option to assemble only those core elements of Office - and similar products - that exactly meet the needs of the people using them. The road to business agility won't run through fat software.

At the recent ITxpo conference in Cannes, research firm Gartner also took a swipe at the size and complexity of enterprise software offerings. Analyst Yvonne Genovese said vendors must break monolithic application suites into smaller, more granular increments so that customers can respond more quickly to changing business needs. Considering that upgrades can take two years to complete, that's not possible today. If vendors break apart the software into smaller pieces, IT can rearrange them more readily as business processes change.

That's the direction SAP is taking with the next generation of R/3, and it's the objective of Microsoft's Project Green, Genovese says. Microsoft is struggling to break apart its Navision, Great Plains, Solomon and Axapta business applications and reassemble the pieces into a cohesive set of components that can be recombined to better suit customers' business needs.

Whether businesses will abandon monolithic enterprise applications and demand more granular, task-specific components in a big way remains to be seen. But increased global competition and the need to react faster to changing market conditions could provide the impetus for IT organisations to begin loudly making such demands. If that happens, Microsoft's decision to enable role-based versions of Longhorn could be just the tip of the iceberg for software vendors.