CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

The greening of Linux

HP open-source chief Martin Fink says Linux can't remain a hobbyist's toy if it's to become a leading operating system.

    For Martin Fink, life is good these days.

    As vice president in charge of Hewlett-Packard's Linux strategy, Fink says HP's decision to indemnify its customers against lawsuits the SCO Group files has been a boon to sales. He also sees HP making inroads with its Linux products against Sun Microsystems' Solaris operating system.

    Fink has been with HP for 18 years, holding a variety of positions in hardware and software support, consulting and telecom sales. He serves as vice president of the board of directors for the , a global consortium of industry leaders dedicated to enabling Linux. He also wrote the book, "The Business and Economics of Linux and Open Source."

    CNET News.com recently caught up with Fink amid his extensive travels to talk about the future of Linux and Unix.

    Q. How much is the SCO suit affecting customers? Is it scaring them away from using Linux? Have you lost any customers as a result?
    A: The reality is that the Linux market overall is growing. What we were seeing was the market continue to grow at 30 percent to 40 percent. At the same time, we were talking to CIOs who were looking at getting Linux deeper into their enterprise, and they were starting to get concerned about the risk.

    Even though a number of them might say SCO doesn't have a case, the fact that they can get sued means they can incur a cost. You have to defend yourself. With the HP indemnification, we said, "We'll take accountability for the solution, and we will defend and take over the costs of the defense." That has had a dramatic effect in terms of the moving customers over the fence and forward with Linux.

    How did the indemnity provision come about?
    We didn't do this on a whim. There was an extensive amount of due diligence. We took an analysis of the risk profile and said we were willing to accept that risk on behalf of our customers. If you look at what some of the others did, IBM and Red Hat countersued. But from a customer's perspective, that didn't solve the problem. The indemnity solved a real problem today. It had a real customer impact today.

    IBM has not indemnified, because it considers the suit groundless. Why even dignify SCO's charges by saying you're going to indemnify against this risk?
    The analyst and customer reaction was positive. The only place I got a couple of negatives was in the open-source community, which had that reaction--that we were dignifying the SCO suit. At the end of the day, when you step into the chief information officer's shoes, he doesn't care if the suit is groundless or if we're dignifying or not dignifying. He cares about, "Will I get sued and incur costs?" What HP is doing is providing protection against those costs. The challenge to IBM: If the suit is groundless, why don't you just indemnify?

    Have you actually won sales as a result of indemnification?
    I'll give you one scenario in the financial-services industry. On the day of our announcement, the CIO of a major company was about to issue a purchase order to IBM. The announcement came out, and he said, "We have to go rethink." The purchase order was blocked purely on the basis of our indemnification.

    Some in the open-source community say the Novell acquisition of SuSE is a sign that the industry is becoming bureaucratized. What's your take?

    The (Linux) community has to recognize that they can't have it both ways
    I consider myself part of the open-source community. The (Linux) community has to recognize that it can't have it both ways. Linux cannot be a hobbyist's toy and be the leading operating platform in the industry at the same time. Those two things are incongruent. For Linux to become a credible part of the enterprise, it has to go through the standard evolution and maturing process. We saw some consolidation. It's MBA 101 here--any new market goes through a whole bunch of little players; then there's consolidation. That's what's happening with Linux. I really like the idea that we now have two strong, powerful players that can compete vigorously.

    How would you characterize Linux sales, in terms of hardware?
    The vast majority is on Intel x86 servers. We're also seeing sales on Itanium, especially for large supercomputing applications. There's an airport whose approach control system is HP Linux on Itanium. Right now, Linux is definitely the leader on Itanium. HP is not all about Linux. We're about multiple operating systems. Our analysis shows that 85 percent of enterprises have multiple operating systems. The idea of saying, "The world is about Linux" is not the real world. The real world is, "You've got to have a strong Linux solution, but you've also got to have Windows and Unix." Our Systems Insight Manager (formerly Nimbus) is a platform that looks across the whole thing.

    HP has a close relationship with Microsoft. How does pushing Linux affect that relationship?
    Our strategy is a multi-operating system strategy. IBM is pushing Linux at the expense of other things. We're pushing Linux as part of a complete enterprise solution. Does Microsoft want HP to be selling Linux stuff? No. But at the same time, they understand that Linux is in the market, and we have to compete. It's not about competing with Microsoft. HP is not poking a finger in their eye. It's about competing with Dell and IBM.

    What are you seeing with Linux on the desktop?
    Linux on the desktop is definitely an area where hype is ahead of reality by orders of magnitude. There's a sexiness around the idea of taking on Microsoft. The reality is that (desktop Linux) is still less than 2 percent of the market, but at the same time, we certify and sell a number of Linux desktop solutions. There are two areas of interest: the engineering desktop with folks like DreamWorks and Disney, and the application developer. We are certifying our notebooks with Linux, and the target there is big-time application developers. In the developing countries that don't have the Windows legacy--like India, China, Asia Pacific and the Eastern bloc--we see some pretty significant volume there.

    What about Sun's announcement at Comdex about its agreement to sell Linux desktops to China?
    I'm in the process of doing a reality check. I was personally in China a couple of months ago. I signed a deal with Red Flag Software, which has 90 percent of the Linux desktops in China. (With Sun's announcement), I'm really confused--is everybody buying two? It didn't add up for me. If I've got the relationship with Red Flag, I have no answer for how Sun could possibly be doing that much business.

    It seems like people have already acclimated to two platforms.
    That's an unknown. We have to wait and see how it plays out. That (Linux) desktop is not mature. You take the Windows desktop, and it's more sophisticated, more developed. We've now got Red Hat saying, "If you're doing a desktop, stick to Windows." Customers say they have to set up a new competency center and support two operating systems. When you think about it, customers have done that for a long time--on the server side. I'm not sure what they'll do on the desktop.

    Are you steering customers from HP-UX to Linux?
    If they're more business application-oriented, the path is HP-UX. Four to five years from now, from a server perspective, Unix is $20 billion, Windows is $20 billion and Linux is $9 billion. That's why we don't do "It's all about Linux." That's not reality. From a migration perspective, we have some 64-bit customers who have moved to Linux. Those who are running huge SAP systems are still HP-UX.

    Do you see a time when Unix will disappear?
    Any planning beyond three years is way out there. In the three-year time frame, which is what we call the foreseeable future, Unix is flat. It's not a growth business. A lot of the growth is going to Linux, but the Unix business is not going away. There are players in the Unix market smack in the sights of Linux. Solaris is the Unix going away.

    Why is Sun the big Linux victim?

    I think we have a personality issue with Mr. McNealy.
    When Linux started, it was mostly on the Web server applications. That was the market where Sun exploded. Sun's business is all skewed into the low-end business. HP and IBM are skewed mid to high. Linux came in at the low end. The cost value proposition and the low switching cost meant it was a dead-center nuclear explosion, where Sun's business was. Everyplace Linux is strong is where Sun historically has been strong and HP has been weak. For us, this is a completely incremental play. It's a beautiful thing.

    Put on the hat of your counterpart at Sun, and tell me what you would do.
    That's a hard question. They have a fundamentally broken business model. Sun is trying to play its cards in too many places. They're defocused. They're making irrational, emotional decisions. With the desktop, that's an emotional thing--trying to take on Bill Gates and stick it in his eye. It is not grounded on any business reality. If you're Dell, and your business is growing, and you want to grow into new areas like consumer electronics, that makes sense. If your business is tanking, and you're on your way down, you should be focusing on your core business and getting your cost envelope to something you can afford, and then say, "How do I grow from here?"

    Audiocast
    arrow Sun believes it has the most complete Linux strategy of any of the top vendors
    play audio

    I think Sun has to paint an end-of-life picture for Sparc. People are internalizing some of their press releases around (Advanced Micro Devices') Opteron as that being the alternative. Having done the migration from (HP's chip) PA-RISC to Itanium, we've been doing that for eight years. You can't just flip a switch and say, "Yesterday it was Sparc, today it's Opteron." It doesn't work like that. You need to get your ISV (independent software vendor) partners there; you need to get your virtualization story straight. The other challenge is that Solaris on Opteron and Solaris on Sparc are two completely different operating systems. You can't just take all your Solaris Sparc stuff and move it to Opteron.

    Does (Sun CEO Scott) McNealy rub people the wrong way?
    I think we have a personality issue with Mr. McNealy. He's being a little too arrogant and cocky for the position he's in. A few spoonfuls of humble pie might be appropriate. But that's the board of directors' job.