Pacific Connection(英語)

Behind the Novell/SUSE Linux Purchase

There was a time earlier in its history when Novell tried to take on Microsoft. It owned DR-DOS, an MS-DOS competitor. It owned WordPerfect, a competitor to Microsoft Word. It patched together an office suite to try and compete with Microsoft Office. None of this worked, of course. Bill Gates slept soundly at night. Now, in a much more limited sense, Novell is taking on Microsoft once more. The battlefield is smaller-Microsoft is a much larger company. And in some ways, the move on Novell's part is more defensive. Like every other company in Microsoft's gunsites, Novell is just trying to stay alive. But behind Novell's purchase of SUSE Linux is the desire for the compete more effectively against a beomuth, using Linux to open doors that might otherwise be shut. The purchase is also an attempt to once and for all re-position Novell as something more than a purveyor of Netware, the network operating system whose marketshare keeps declining. Novell has long been much more than Netware. With the purchase of SUSE Linux, it hopes that the rest of the world will finally agree.

Novell purchased SUSE Linux purchase last November for $210 million. Doing so got the company a lot of attention. For the Linux community, the acquisition meant that a company with a proven track record in enterprise software now controlled the second largest Linux distribution. For Novell, the purchase put the spotlight on its Linux activity, which had been slowing building since 2002. The company announced the puchase by trumpeting its full Linux stack, supported by security, networking, collaboration and management tools.

In a statement, Novell chairman and CEO Jack Messman noted that Novell had offered cross-platform products for the past four yeasrs with Linux an increasingly important part of that strategy. He sait said Novell would be the only billion-dollar software company with a Linux distribution and would maintain a worldwide technical staff of more than 600 to support it. IBM, the most prominent Linux booster and already an investor in Red Hat, agreed-investing $50 million in Novell (in convertible preferred stock). So did Dan Kusnetzky, vice president of system software research for the research firm IDC, who told Infoworld magazine that "Linux now has a legitimate Pepsi to Red Hat's Coke." He noted that Novell has a chance to convert Red Hat users, some of whom are unhappy about that company's licensing model. Another analyst, John Enck of Gartner, said that Novell was now in a position to go head-to-head at the OS level with Microsoft. Novell wouldn't quite want to go that far-it can be a dangerous business going head-to-head with Microsoft on anything. But the fact of the matter is that, SUSE Linux in hand, Novell can now knock on doors and get the attention of companies who might otherwise have gone to Microsoft.

Novell also continues to insist that Linux will not replace Netware. But Linux appears very much to be Netware's successor. The company's surveys show that about a third of its customers leaving NetWare move Linux, with half of those remaining considering the move down the road. Those customers are hesitating because of concerns over vendor support. Novell can now claim it is the only $1 billion-plus software company with a Linux distribution, with a worldwide infrastructure to support it. In other words, if you are going to drop Netware, Novell would still like to be your OS vendor of choice.

For Novell-a change of omage and a foot in the door

For IDC analyst Dan Kusnetzky, the Novell SUSE purchase is the capstone of a strategy that dates back to June 2002, when Novell's purchased of SilverStream Software. "SilverStream makes tools to encapsulate and integrate applications," Kusnetzky says. "The assumption is that there won't be a lot more new enterprise software written. More likely, the existing packages will be integratable-not necessarily from the bottom up, but encapsulating them, filtering the input and output, and linking them together with higher levels of software."

The bigger move came in August 2003 with Novell's aquisitioin of Ximian. "Ximian makes user interface software that lets a Windows or Macintosh user sit down at a Linux or Unix system and and be proeuctive. It hides the lower level protocols, allowing a Linux server or desktop to operate similarly to a more familiar machine." Other products include a desktop electronic messaging client that allows a Linux server to co-exist in a Microsoft Exchange Server world (Microsoft Exchange Server handles email and other messeaging) and a version of the OpenOffice desktop office application.

Perhaps Ximian's best known product is Red Carpet, which installs, updates and managing software on Linux workstations. "Red Carpet gives you a working system without your people having to be experts in multiple streams of open system development and how they interact with each other" Kusnetzky says. A newer Ximian package is Mono, which offers "C++ development libraries that allow Linux to be a client or server in a world that is using the .Net architecture."

Kusnetzky looks at all these packages and thinks Novell saw some mapping between them and its own products. Ximian's asset management with Red Carpet fits well with the ZENworks portfolio. The desktop client works well with Novell's GroupWise comunication and collaboration tool and Netmail email and calandering system, and the Mono project in conjunction with SilverStream's technology enhances Novell's role as the glue that ties the organization together. "Novell has a powerful metadirectory product, with a whole series of assset, system, patch and security managmenet pieces built atop of it, as well as Netmail and GroupWise." All of this supports Linux, among other operating systems. Indeed, the only missing link for Novell to offer a complete top-to-bottom stack was a Linux distribution.

And why is a complete Linux stack neeeded? "Because it's pretty clear that opportunities with Microsoft are going to be problematic going forward-because Microsoft is increasingly putting these tools in the operating system," Kusnetzky says. "So that if you are buying Windows Server 2003 and you want identity management, you are going to use Active Directory, because nothing else works. You may be able to link ActiveDirectory with a metadirectory service, but you will still be using ActiveDirectory on Windows. With the Linux acquisition, Novell at least has the chance to discuss the rest of its stack of software in places that might not have considered them before."

Number two-but with more clout

Even with Novell behind it, SUSE Linux is still the number two distribution behind Red Hat. That's still a pretty good feat considering that Kusnetzky's group at IDC actually tracks 147 different Linux suppliers, many of whom are not known outside their region. "But now the number two distrubution has behind it a company with a $4 billion market cap, who is a worldwise enterprise player with alliances with all the major hardware suppliers and the majority of major software suppliers," Kusnetzky says. "Novell is saying: we have all the tools top to bottom to help you integrate your systems, manage them effectively remotely or locally, based on an identity management ssytem that is powerful, yet easy to use and has been deployed all over the world."

Unlike Red Hat, Novell company receives revenues from many different sources-not just Linux. "If Novell wanted to, it could drive the price of Linux way down and depend on revenue from other sources to make Linux more enticing," Kusnetzky says. He thinks that Novell has a chance to attract Red Hat users who are unhappy with that company's licensing model. "Novell can say to prospective customers: "we're offering the same class of enterprise product as Red Hat, with the same level of service-but we'll give it to you the way you want. You don't have to automatically buy the enterprise class license if you are planning, for example, to buy the software once and deploy it on a 2000 different nodes in a compute cluster." Kusnetzky also thinks Novell has done a better job than Red Hat of inemnifying Linux customers against a lawsuit from SCO.

Kusnetzky does have two reservations about how Novell is proceeding. One has to do with what he thinks is an entrenched corporate culture. In purchasing Ximian and SUSE, Novell has acquired two companies that "live in Internet time." Those companies understand that product lifecycles that used to measure a year or two have now shrunk. "In the world of the Internet, you've got four to five months to develop a product. You bring it out, it lives for six months, and it's replaced by the next one. You have to be constantly seen or people forget you." Novell, he says, has made beautifully engineered products, but not made this critical transition. "SUSE was very good in acting in Internet time, but the people who did that have been leaking away. And the old Novell people are now being pressed into service to speak to business people marching to the Internet drum."

Kusnetzky also notes that this story of how SUSE works with other Novell acquisitions--the one he presents here--is harder to get from Novell itself. "If you search on the Web, you could piece this story together-but how many people are persistant enough to do that?" He contends that the roadmap story on Novell's own website is so fragmented that a cohesive strategy doesn't come forward.

So here's a chance for Novell a chance to speak for an interview with Jeff Hawkins, Novell's Jeff Hawkins, vice president of product management and the company's chief Linex strategist.

What is Novell's Linux strategy. Given your acquisitions, what are you trying to do?

You have to look at where Novell was in the marketplace. We made a couple of acquisitions that positioned us in the Web services area, but our growth was flat and declining, and a lot of that growth was connected to the Netware operating system, As Netware came under attack from both Microsoft and open source solutions, we felt that we had to make a significant bet in the market- and that led us to the SUSE and Ximian acquisitions. That, in turn, gave us a platform on top of which we could then sell a lot of our other value-based services-things like identity-based services--managing the identities of people within the organization. With Linux and these open source solutions, we were seen as having solutions that extended beyond Netware, reaching to the Windows operating environment, as well.

Most if not all of your pieces already ran on Netware, Windows and several flavors of Unix and Linux. Why buy SUSE Linux?

We successfully engineered our products to support many of these platforms. But we had a gap in getting customers to perceive that we had those solutions and could sell them into those market spaces. This acquisition has broken through that gap in ways that engineering alone could not.

When you finally got your own distribution, was that the final piece of the puzzle?

A lot was the market momentum was around Linux itself. It really is a lightning rod for the world in software right now. so acquiring SUSE really jolted everyone that watches this industry. It was significant.

So the act of purchasing may have been as important in terms of perception as in anything technological?

Owning the distribution means that you can be a complete solution provider for your entire service pack. Consider the problem we faced from the sales perspective. We would go into a customer and say: if you want an identity-based solution, you must first buy this or that operating system, then come back and talk to us. But those other vendors may not support putting our solution on their stack. They'll try to sell their own managed space solutions, instead.

Are you referring largely to Microsoft?

Microsoft, yes. Not exclusively, but they clearly were and are our largest competitor.

How does your approach now differ Red Hat's?

I believe that Novell goes beyond just a distribution and has a significant number of value-added services that you can't get from Red Hat. Clearly our management with our ZENWorks product portfolio is far superior than anything you could get from Red Hat. With the Xinian acquisition, we got Red Carpet, a world class distribution engine for managing Linux deployments. We have identitiy services, Web services, and the other things that sit above just the operating system-that you can't get that from Red Hat. So if all you want is an operating system and the set of open source solutions that come with that operating system, then Red Hat is a viable choice. But customers have other needs that Novell believes we can fill. We have a worldwide presence with support centers throughout the world. We are a significant education and training vendor with consultants throughout the world. We provide the complete ecosystem that is required to foster worldwide growth.

Do you have any special considerations in terms of your business in Japan?

Only that we are intrigued with the potential of Linux within Japan and Asia Pacific in general. The region is embracing open source and Linux as much as any other part of the world, and is in fact leading in a number of cases.

You are on the OSDL [Open Source Development Labs] board. Where do you see Linux headed these days?

I see there being a number of potential areas of growth for Linux. They include the desktop, workgroups using server-based workloads to collaborate, and the data center. High performance computing is another, typically for doing computationally intensive business-related research. There is also the embedded Linux market and the carrier, or telco market. I see Linux advancing in every one of these spheres.

Is it fair to say that the desktop is the most difficult for Linux to break in?

Yes, it's the most difficult to crack because of what it takes for users to move over. But the remarkable thing about Linux is that it is unlike any competitor that Microsoft has ever faced. You can't simply eliminate it through some kind of a tactical strike. It's there; it keeps advancing, getting better. It keeps innovating.

The other thing about the desktop is that the general knowledge worker is the last mountain to be scaled. But there's are other people using client machines-from the guy on the shop floor to the help desk technician. There are airport information kiosks and point-of-sales terminals. All those sites are already perfect candidates for a Linux desktop. We're seeing a slow adoption and the industry embracing it over time.

What's running on Novell desktops?

By the end of July, every employee will be using OpenOffice as their office package. By the end of October we'll have roughly overy 50 percent of the desktops at Novell running on Linux.

Where do you see Novell headed?

I see Novell fully embracing open source and being a partner with the open source community. I see us living happily in a mixed source environment where open source provides a great foundation and a great community for getting traction and collaboration on a global scale. I see us building value that sits on top of what's unique to Novell. I see us really focusing on identity-based solutions, things that tie who you are to the things that you want to do.

We really manage the entire life cycle of the IT assets of an organization, from the time it shows up on the dock to the time it is retired-installing the right applications, managing user profiles, updating software and patching. Add to that our commitment to open source, and we bring even more to the market.

Dan Kusnetzky said that Novell has not been used to working at Internet time, and he sees some loss of key personnel from SUSE. Is that a fair assessment?

We've been able to retain the staffs of these two very important organizations. And one of the purposes of the acquistions was to change the DNA of Novell, not the other way around. The development culture and methodologies that are used at both Ximian and SUSE have been infused into Novell and the teams have been integrated in some very interesting ways.

But to build a quality software product that has 5 millions lines takes a lot of time. That's true regardless of how or where you are doing it. They way to get on Internet time is to break down the dependencies between code and build more stabilized interfaces, and then create packages and teams that are smaller in size and smaller in engineers. There's no magic-only a way of working more intelligently.