Without question, open source is one of the most important forces in enterprise IT. Apache is still the reigning champion of Web server software. Red Hat and MySQL are two of the most-watched software vendors today. And the success of Linux is such that even Microsoft, once considered the unstoppable juggernaut of the OS market, is being forced to play defense.

The open source model is often said to fly in the face of traditional commercial software development. But now Novell, the former undisputed leader in proprietary networking software, is banking on Linux and open source to rescue its business from years of decline. Where once it seemed fated for obscurity, Novell is determined to regain its former status by transforming itself into one of the industry's leading open source vendors.

Novell first aligned itself with the open source community when it launched its collaborative development site, forge.novell.com, in early 2003. But it wasn't until Novell purchased open source desktop software vendor Ximian in August 2003 that the company's new direction began to solidify in the public eye. Then in January 2004, Novell acquired SuSE, the leading Linux vendor in Europe and the No. 2 player worldwide, after Red Hat.

But even though it seems to be holding all the right cards, Novell faces tough odds. In recent years, tough competition from Microsoft and dwindling support from third-party developers have caused Novell's once-loyal base to look elsewhere for infrastructure needs. Unless it can win back the loyalty of the industry, Novell's new, Linux-centric message could fall on deaf ears.

LAN leader dethroned

Novell's fortunes weren't always so dire. At one time, its name was synonymous with PC networking. The debut of its NetWare file sharing technology in 1983 came as a revelation. Previously, networking had been an arcane affair, used for connecting workstations to expensive mainframes and minicomputers. With NetWare, any workgroup could share files and printers using straightforward, native PC software.

Throughout the 1990s, Novell built on NetWare's momentum, releasing technically superior products such as NDS -- now known as eDirectory -- and the ZENworks configuration management system. Analysts estimate that at the height of its success Novell commanded as much as 75 percent of the PC networking market. But as the years wore on, its star dimmed.

The rise of the Internet and increasing dominance of Microsoft products such as Windows networking and Active Directory -- among other factors -- steadily eroded Novell's customer base. Seemingly blind to the threat to its core business, Novell made a series of missteps, including the acquisition of WordPerfect and the Quattro Pro database in a bid for the desktop-applications market. A revolving executive team and an ongoing brain drain of developers to the competition compounded its woes.

Today, left with less than a 4 percent market share, the former LAN leader teeters on the brink of irrelevance. At Novell's darkest hour, open source could be the company's last, best hope. And yet, company executives resist being labeled.

Back to the source

"No, it's not fair to characterize us as an open source company, and it's not fair to characterize us as a closed source company," says Novell CTO Alan Nugent. "I don't know that there's any particular name for it, but I guess you could call us a 'mixed-source company,' at this point."

Novell's strategy seeks to leverage the strengths of both the open source community and its own portfolio of proprietary technologies. To that end, it has begun porting its entire application stack -- including eDirectory and its identity-management and Web services technologies -- to Linux. By the end of this year, Novell plans to release Open Enterprise Server, which will be the culmination of its efforts to create a unified code base of Novell services for either the Linux or the NetWare kernel.

"Our stack and our vision, strategically, really haven't changed dramatically as a result of the acquisitions. Yes, we've embellished a little bit, but the vision remains the same," Nugent says. For Novell customers, that's both good and bad. Although there has never been much debate about the technical merits of Novell's products, those applications alone won't be enough to win customers over to a Linux-based platform.

"You don't hear complaining about NetWare 6; you don't hear complaining about ZENworks," says Laura DiDio, an analyst at The Yankee Group and a longtime Novell-watcher. "Certainly eDirectory, from a technical perspective, is still more advanced than Active Directory. The issue, though, is that as Novell's fortunes have waned, the developer community has pretty much abandoned them."

Developers are essential

Gary Hein, vice president of application platform strategies at Burton Group and a former Novell executive, agrees. "Novell was famous for stability and performance and infamous for developers and applications. Especially NetWare today; you can barely get a virus checker and a backup package for the NetWare server."

The stagnation of the NetWare platform has prompted many longtime Novell customers to jump ship to the competition and may be the strongest argument for the company's new direction. With the shift in focus to Linux, Novell gains a huge developer base and not just from the open source community. Commercial vendors that support their products on Linux become selling points as well. "You can now go back and run Oracle on a Novell platform, which you haven't been able to do for eight, nine, 10 years," Hein says.

Ironically, this infusion of developers could actually have a chilling effect in the short term. The perception that Novell is abandoning the NetWare OS in favor of Linux could accelerate the erosion of its customer base, particularly among companies not ready to gamble on open source.

This is a scenario Novell cannot afford. But here is where Linux could play a crucial role. Enterprise networking customers may be more likely to consider offerings from Novell if the buy-in cost doesn't include supporting a niche OS such as NetWare. More importantly, the technologies acquired from Ximian and SuSE put Novell in a position to go after an entirely new market: the desktop.

Capturing new markets

"We are in the desktop Linux business; there's no two ways around it," says Novell's Nugent. Whereas there may not be much profit in selling Linux desktops to the home-enthusiast market, Nugent sees real potential in the corporate sector.

"That chunk of the market is speaking very loudly," Nugent says. "And what they're saying is, if Linux can get to the point where it is a viable, lower-cost alternative and roughly equivalent in terms of capabilities -- (for example,) a word processor that behaves a lot like Microsoft Word or WordPerfect or whatever -- then they're happy with that."

Unfortunately, Novell's chief competition on both the server and the desktop is Microsoft. Convincing IT shops to switch from Windows to Linux is a hard sell, especially given that open source office suites are available for Windows and that hard figures about switching to Linux desktops -- including training and systems management costs -- are elusive. At the end of the day, those costs could easily outstrip any savings on the OS software itself.

So, although it seems clear that a market for Linux and open source in the enterprise exists, Novell must find a way to capitalize on it. SuSE and Ximian were good choices, but Novell has acquired companies before. It bought rights to the Unix operating system from AT&T in 1993, only to sell them two years later. It picked up WordPerfect and Quattro Pro in the mid-90s but never made a splash in desktop applications. And although it acquired Web services vendor SilverStream in 2002, it still hasn't emerged as a serious contender in that market.

The greatest beneficiaries of Novell's recent moves may be its open source acquisitions themselves. "I'm a happy employee," says Ximian CTO Miguel de Icaza. "Novell already has a massive distribution channel, not only in the (United States) but also in the world. All of a sudden going from getting one big customer to having a list of 20 people who want to deploy -- that's a big change." Whether it's a change that translates to profits for Novell, however, remains to be seen.

It's all about execution

Despite the challenges facing Novell, Burton's Hein is optimistic. "I'm glad to see that they've consolidated their desktop strategy under (Ximian's) Nat Friedman," he says. "That, to me, is a positive sign. Now, it's one thing to have your reporting structure on paper; it's another thing to see how this thing actually pans out."

Indeed, Novell faces the daunting task of integrating its new open source divisions with its core business practice. So far, despite SuSE and Ximian press releases issued under Novell letterhead, there's been little to indicate a comprehensive and coherent strategy that unifies the open source culture with Novell's corporate one.

The Yankee Group's DiDio fears that, rather than solidifying its product direction, Novell has devoted too much energy to side issues, including a costly legal battle with The SCO Group over ownership of intellectual property related to the Unix OS.

Time may be running out. The latest research from The Yankee Group suggests that eight out of 10 businesses will undergo a major systems upgrade in the 2004-2005 time frame. If Novell doesn't produce a full-fledged strategy soon, it could mean countless missed opportunities.

"Even stalwart MIS departments, which would love to go from NetWare to SuSE Linux, are not going to be able to convince their CxOs to commit to the Novell platform until there is a fully fleshed-out product road map, both tactical and strategic direction, and a full list of what everything is going to cost," DiDio says. "Is this Novell's best chance in a while? Yeah, but I've seen them blow so many good chances. It looks good on paper, but they have to execute."

The end of the line for NetWare?

To migrate or not to migrate? That is the question facing NetWare administrators, as IT departments begin planning their first major upgrades since overhauling their systems for Y2K. The message from Novell is clear: Linux is the future, with NetWare likely relegated to legacy status.

So, what does that mean for the estimated 90 million users and 4 million servers worldwide that rely on Novell's NetWare operating system?

"The short answer for what it means to the NetWare base is absolutely nothing if they want it to mean absolutely nothing," says Novell CTO Alan Nugent. "And it can mean a lot to them if they want it to."

Nugent insists that Novell will continue to support NetWare customers as long as it has customers to support, citing the platform's continued support for new hardware such as SATA drives and blade servers.

But for Novell, part of the business case for open source involves minimizing direct investment in development. For example, Novell plans to migrate its J2EE application server customers from its own exteNd product to the open source JBoss. Given this strategy, the idea that Novell would pour significant resources into maintaining NetWare seems far-fetched.

"They have a huge installed base, and I've got to tell you, I think NetWare is very cool," says Gary Hein, vice president of application platform strategies at Burton Group. "There are a lot of things in NetWare that you still can't do on Windows or Linux. But at the end of the day, it's not about how cool technology is. I think the writing is on the wall, and NetWare is going to be in continued decline."

If NetWare is perceived as a dead end, the challenge facing Novell is to convince customers who do decide to migrate that Linux is the best choice. If Novell fails, it stands to lose many of those customers to the Windows platform.

But winning customers over to Linux is likely to be a tough haul. Many, particularly those in vertical industries, doubt that Linux can meet their needs, according to Laura DiDio, an analyst at The Yankee Group.

"Time and time again, particularly in health care and legal (industries), people are saying, 'Forget it. We will never go to Linux because our core, mission-critical health-care applications are not supported on Linux,' " DiDio explains.

Unless Novell can remedy its Achilles' heel -- a lack of developer support -- the decision to go with a Linux or NetWare kernel may be moot.