The fair-weather "friends of Linux"

The fair-weather "friends of Linux"

Summary: Ever since Microsoft entered into their agreement with Novell, everyone has been speculating that Microsoft will sue RedHat. My question is ... Why would they bother?


I just finished reading John Carroll's article Why Microsoft won’t assault Linux and while I agree with his conclusion, I don't agree with his reasoning.  John opens up suggesting that a "full legal assault" on Linux would be self-defeating.  I don't think so but I do think that filing suit against RedHat would be an unnecessary waste of Microsoft's time and money. 

I would like to dispense with his first point: Patents cut both ways.  While this is certainly true, John assumes that there are powerful "friends of Linux" who would counter-sue Microsoft in a heartbeat if Microsoft were to use it's patents against RedHat and he reminds us that: Nobody has more patents than IBM.  Does John think IBM would sue Microsoft to save Linux?  I don't think so...

This argument assumes that IBM cares whether Linux lives or dies.  Why would IBM care?  IBM sells Linux because a significant number of it's customers insist on buying Linux.  If IBMs customers were to suddenly decide that investing in Linux were too much of a risk, IBM would happily sell them AIX instead (running on proprietary hardware, no less). 

The other major patent holders, Hewlett-Packard & Sun Microsystems, also have substantial UNIX assets which they would prefer to sell instead of Linux.  In short, all the major players sell Linux to their customers because their customers insist upon it.  Not because they are "friends of Linux". 

John's second point is that Microsoft would suffer long-lasting ill will.  Well this is also true but let's face it, no one in the Linux camp has many positive things to say about Microsoft anyway. 

John's third and final point, regulatory oversight, is the most compelling.  But I don't see this as an intellectual property issue.  In other words, his point is correct but not necessarily his argument. 

Microsoft has been declared a monopoly in both the U.S.A. and in the E.U.  They have been found guilty of numerous anti-competitive practices and they find themselves under immense scrutiny for that reason.  Still, despite regulatory efforts, Microsoft continues to dominate the desktop space for a variety of reasons -- not the least of which is that neither Apple nor any of the other top-tier vendors in that space care to compete for commodity dollars.   

Since porting to Intel, Macintosh fans are quick to point at that a similarly-equipped Dell is no less expensive than a Macintosh.  If Apple chose to, they could configure an entry-level Macintosh at Dell price-points.  Since Apple makes their margins on hardware though, selling commodity desktops is not in their best interest. 

Similarly, Sun has sold an Intel port of Solaris, its UNIX offering, for years and even briefly marketed a Linux-based Java Desktop System.  Today, Sun even gives away Solaris 10 for Intel but the consumer cannot buy it preloaded on competitively priced Intel hardware.  Like Apple, Sun is in the hardware business -- and also the services business -- and it's not in their interest to compete in the commodity desktop space either.  It's no different with IBM and HP -- though HP is in a position to market pre-loaded Linux on it's Intel workstation line, if it so chose.

On the Linux side, there are LOTS of choices.  And all will run on a wide variety of Intel hardware, available at commodity prices.  So why aren't there more Linux desktops out there?  Because none of the top-tier hardware vendors offer a workstation to the consumer with Linux pre-loaded.  There are a small handful of hardware vendors offering Linspire pre-loaded but these vendors are pretty much unheard of. 

Geeks like us have lots of desktop choices:  We could chose Windows, Macintosh, Solaris 10, even UnixWare from (ugh) SCO.  Or we can choose any number of Linux distributions, from RHEL to SLED, from Ubuntu to Slackware, to Linspire or anything in between, and put it on any Intel hardware we wish.  But what about the consumer -- who knows little, if anything, about operating systems and who wants his OS preloaded and wants his hardware from a company he's heard of? 

Windows or Macintosh are his only choices, and the price gap makes a compelling argument in favor of Windows.

You might ask ... "Then, why does Microsoft even care about Linux at all?"

Microsoft cares about Linux for the same reason it cares about UNIX:  The real profits to be made in this industry are in the enterprise machine room -- selling servers.  In this environment, Windows competes on a level playing field with UNIX and Linux.  That means they compete daily with IBM, HP, Sun, and of course, RedHat.  Novell and SCO compete in that space as well but they are not big players.  Nor are the other Linux vendors.   

So, if IBM, Sun, and HP care little about the future of Linux, why wouldn't Microsoft challenge RedHat in court?  Because they don't have to!

In order to keep RedHat at bay, all Microsoft has to do is create that all important Fear, Uncertainty, and Doubt (FUD) in the minds of would-be RHEL customers.  If customers know that Microsoft already has established cross-patent agreements with Sun and Novell, and if there is "patent parity" (a kind of a 'balance of power') between Microsoft and its other competitors in the space (IBM & HP), but RedHat while offering "indemnification" of sorts, has no such protections from serious litigation, Microsoft need not follow through with it's implied threats. 

Sure, the cross-patent agreement gives Novell a leg-up in the space but they are a much smaller player than RedHat.  Microsoft can afford to throw them a bone if the end result is that RedHat gets squeezed out of the enterprise server market. 

Don't expect Microsoft to go after RedHat anytime soon.  For now, the threat of litigation ought to give RedHat customers sufficient pause to keep Microsoft happy.  At least until the verdicts are in regarding the SCO suit against IBM and the SCO-Novell litigation over UNIX intellectual property rights.  The outcomes of those trials will have a much greater effect on the future of Linux in the machine room (and hence on RedHat) than anything Microsoft is liable to do. 

As for Linux on the desktop, it's up to the Linux vendors themselves to make that happen.  Will they?  Only time will tell.

Topic: Operating Systems

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Dear Mr wagner PLease explain

    If microsoft is more intrest where the big bucks are ( into the buissness (typo) server and service soft )

    So why dont microsoft dont make a free or low cost OS ......

    I do think that MS is so afraid that the more home user use linux at home . The next generation of user will bring it into class ,work and all facet of life .......

    I do think that MS is very afraid of a other anti-trust lawsuit and they with the DEAL with novel will provide them with the free OS that they need .

    thx you sorry for the english
    • No money in it.

      "So why dont microsoft dont make a free or low cost OS ......"

      No need to, they are replacing Unix faster than Linux is, they now hold serious cards in Linux. No reason to go the cheap route to beat Linux.
    • Microsoft makes more money ...

      ... if you buy new hardware than if you buy a shrink-wrapped version of Windows -- even at exhorbitant prices. That's because they can send ONE copy to Dell and sell millions of licenses.

      In short, they don't WANT you to buy a shrink-wrapped copy of Windows.
      M Wagner
  • You forgot about the BSDs

    I think these deserve a mention, don't you?
    • MacOSX is the most significant ...

      ... example of the BSDs still in the market place. Everyone else competing for enterprise (or even consumer) dollars is running a derivation of SVR4 or Linux. The BSDs are just not big players.
      M Wagner
  • Real profits vs. ... what? Fake profits?

    "The real profits to be made in this industry are in the enterprise machine room ? selling servers."

    Oh really?

    As of Nov 3, Microsoft's market cap was $282.43 billion. Cisco comes in at $144.29 billion, Google at $143.6 billion, IBM at $137.7 billion, and Intel at $118.3 billion.

    MS has a market cap twice as large as it's nearest competitor in the industry (defining the industry pretty broadly).

    But the "real" money is in the server space. All that other money MS made isn't really "real."

    Yes, margins in the server space may be higher, but the commodity space is larger -- and MS owns it. To say that Apple doesn't want to play in the commodity space is like saying that I don't really want to win the lottery.

    Everybody wants to play in the commodity space. They just don't win.

    And what happens when MS commodifies the server market? : )

    The really big money is always in commodities.
    • The last thing

      [i]And what happens when MS commodifies the server market? : )[/i]

      That's the last thing that Microsoft wants -- you can't make 80% margins in a commodity space.

      Linux, now -- there's a possible route to commodity server software. That's also the reason that Microsoft has to get rid of Linux -- it keeps a lid on MS' margins.

      Of course, having the desktop commoditized would also do Bad Things to Microsoft's market cap. If you look at their market cap/sales ratio, it's several times that of the rest of the companies on your list. No way is MS going to increase their volume enough to make up for having their margins dragged down to commodity levels.
      Yagotta B. Kidding
    • Microsoft gained dominance ...

      ... long before the PC became a commodity. IBM got out of the PC business precisely because it became a commodity. If you have the volume, you can play in the commodity marketplace.

      If you don't -- you have to live with loses while you build up your volume through superior products or superior marketing.

      Today, MS is in a position to compete with the UNIX vendors in the enterprise precisely because their commodity market dominance is so high. They didn't get there by accident.

      Microsoft is patient -- they started going after the server market a decade ago, when UNIX was their main target, and now they are a player.

      None of their biggest competitors care about Linux, except RedHat -- who is a thorn in everybody's side -- precisely because they have an excellent product and the support structure to compete in the lucrative server space.

      If Apple wantsed to 'take a loss and make it up on volume' they could sell a $400 desktop just like Dell but they don't want that business -- and neither does RedHat, or Sun, or IBM.

      Your market cap figures demonstrate my point. Microsoft's market cap is higher than IBM's, thanks largely to their investment in 'bricks and mortar' to house their staff. But, IBM is still the largest IT company in the world in revenues, where it counts. MS improves their position by competeing in the same space as IBM. But IBM does not improve its market position by competing on the commodity desktop.
      M Wagner
  • IBM will protect Linux

    The point David makes on IBM would be just as happy to sell AIX instead of Linux, does not hold weight, for 3 reasons:

    1. AIX does not run on Intel. It's highly tuned to run only on IBM's big iron servers using IBM's Power cpus. IBM needs Linux, or something else that runs on Intel, to run a large portion of their smaller and/or mid sized servers. They could conceivably move to BSD or Solaris, or even port their own AIX. But BSD does not have the large scale hd and software support Linux does, Solaris would probably involve licensing to Sun (an HD competitor), and porting AIX to intel would be a major undertaking.

    2. IBM has a very large, lucrative services business based on Linux (among several other things). Make Linux go "poof", and at the same time make a large chunk of IBM global services go "poof".

    3. Customers demand Linux. Customers know Linux now, and are comfortable with it. And existing customer currently using Linux would not want to have to migrate to a BSD, Solaris, or AIX. There would be no real benefit to them for the areas Linux is being deployed at, and there would be huge migration costs and headaches. IBM will protect it's customers, which means protecting Linux.

    All that said, these things don't necessarily protect Red Hat, they protect Linux in General, as far as IBM is concerned. If Red Hat sinks, IBM can go with another Linux distro (hello, Mr Shuttleworth), or put out their own (even based on RH code, using probably CentOS as a base).

    BTW - all of these things apply to HP as well.
    • Red Hat is NOT Linux

      You need to get your head around that concept. When push comes to shove, IBM will simpoly switch to Novell, after all they do have a major investment in Novell.
      • That's what I said, silly

        I said IBM will protect Linux being used in it's business and by it's customers. I won't necessarily protect Red Hat. IBM could easily come out with "Blue Linux", based on CentOS or something, while Red Hat goes away (if that happens).

        IBM could go with Novell, but I doubt they want to be wrapped in the MS blanket. They are, after all, a fierce competitor of MS.
    • Your number 3 is precisely the point ...

      ... if customers didn't DEMAND Linux, IBM wouldn't sell it. If MS creates enough FUD, RHELinux customers may abandon it on their own out of fear that MS 'might' sue RedHat over IP infringement.

      Even if Novell takes up some of the slack in the Linux market, they cannot compete with IBM's service and would probably even cross-license SUSE if IBM asked them to.

      If SCO wins it's case against IBM (not likely), it could cost IBM hundred of millions (which they HAVE) to get its AIX license restored (probably by acquiring SCO), but it could cost Linux vendors millions (which they don't have) to clean up the infringing code. (If the courts determine it is even possible.)

      BTW, IBM would port AIX to Intel before they would license Solaris (think 'cold day in hell'). The whole point of the UNIX model is that it is easy to cross-compile to any archtecture. The real issue with AIX is that it is heavily laden with journaling features derived from MVS in order to support rock-sold data integrity. These features are too demanding to run on all but the most powerful Intel hardware and might have to be removed to port.

      IBM is not dependent in any way on RedHat and IBM can download Linux for free just like you can. They have all the expertise they need to support Linux but they have UNIX assets they need to recoup. That's why IBM would rather NOT sell Linux.

      As for number 2, exactly how long do you think it would take for IBM to do a 'global replace' on "Linux" to read "AIX"?
      M Wagner
  • SCOX??

    [i]At least until the verdicts are in regarding the SCO suit against IBM and the SCO-Novell litigation over UNIX intellectual property rights. [/i]

    The SCOX situation, despite its boundless hyperbole and capacity for idle amusement, has absolutely nothing to do with patents, and by now nothing to do with copyright either.

    The most that SCOX can get now (according to their own complaint before the Court) is a particularly insane reading of a contract claim. It has no bearing on anyone else at all.
    Yagotta B. Kidding
    • I think that your assessment of SCO's ...

      ... chances are right on the money. And apprently IBM agrees with you or they would have bought SCO hook-line-and-sinker just to get them to stop whining.

      Still, there is a non-zero chance that the courts will find in SCO's favor. It could cost IBM some serious cash to get their AIX license restored but the hit Linux could take would pay them back many times over.

      The key to the SCO-Novell rift is indeed in the reading of the contract. If Novell prevails, they join Sun as holding undisputed rights to the SVR4 codebase. Plus, both vendors have patent protection from Microsoft. This is an enviable position for them to be in when competing with IBM, who holds a UNIX license under Novell's (currently SCO's) oversight. If SCO also retains its rights to UNIX, IBM would proably buy out SCO in order to be on an equal footing with Sun and Novell.

      If it's not IBM, whomever ends up owning SCO assets will be in a most interesting bargaining position.
      M Wagner
  • Here's why IBM cares about Linux: IT MAKES THEM MONEY.

    It never ceases to amaze me -- really, truly amaze me -- that smart folks don't understand what IBM sees in Linux.

    The article quotes the following passage (from an unclear source):

    [i]IBM sells Linux because a significant number of it's customers insist on buying Linux. If IBMs customers were to suddenly decide that investing in Linux were too much of a risk, IBM would happily sell them AIX instead (running on proprietary hardware, no less). [/i]


    IBM sells Linux because it [b]costs them nothing to make[/B]. Well -- [i]almost[/i] nothing.

    This is the big hairy scary truth about Open Source that f/oss advocates for the most part either don't or won't get: [b]Open Source amplifies the disproportionate advantage held by large players, if the large players understand how to "work" the open-source process.[/b]

    Right now, no one understands that better than IBM. Not even Google. Linux is hugely profitable to them. They're making money they'd never have a chance to see if they were only selling AIX. Sure, they enjoy selling proprietary software. But it's not where they make most of their money. [b]They make most of their money selling integration and services[/b]. Linux and other f/oss software is [i]fantastic[/i] for driving services revenue, as long as you make sure that you offer actual -- or at least consistently perceived -- value for the services dollar. IBM offers at least the strong perception of value. They do that because of their tight integration with the f/oss movement.
    • Not as much as AIX would!

      IBM cannot reap the kind of margins out of Linux that it could out of AIX if Linux were not competing against them. If Linux were NO MORE, IBM would happily go about it's merry way selling AIX. Without Linux, RedHat has nothing to sell!

      Linux is NOT FREE to IBM or to any other Linux vendor. In order to make THEIR distribution more attractive than the other guy's distribution, they have to ADD VALUE to it. But wait, Linux customers will nto pay the high margins that AIX customers will pay so while IBM is making money out of Linux, they could make more if they didn't have to compete with Linux.

      IBM pays not royalties to sell AIX. They haven't for years so AIX costs them only what they want to spend on it.

      Yes, their bread and butter is services. Same with RedHat but without something to SERVICE, RedHat would be dead. IBM could live on its patent portfolio alone. And they will be servicing their ahrdware and AIX long after the founders of RedHat are dead and gone.

      Linux is good for the industry because of the competition it creates but neither IBM nor any of its biggest competitiors cares whther Linux lives of dies. That said, the only thing that is likely to kill of Linux is a wildly success patent infringement suit targetted at it. It only other vulnerbaility is the GPL itself -- which could be rewitten if need be.
      M Wagner
      • Your thinking is dated

        You're still thinking in old models. Open source does something very very valuable for IBM: It displaces their costs onto other entities (e.g., universities, other corporations, private hobbyists).

        Their investment is small: A few FTEs assigned to work on a f/oss project that's important to IBM gives them a powerful voice at the table for that particular project. They can essentially steer it where they want it to go.

        Proprietary software has a much higher carrying cost for the company. And with IBM's traditional pricing models, they'd be driving lower-end, more cost-conscious customers to other vendors. Now, with f/oss (and in particular with Linux), they can sell services and integration to those customers. I repeat: they would not have access to that market without Linux.

        Would some dinosaurs at IBM love it if we were back in the '90s and Linux didn't matter commercially? Sure. They're idiots. They forget that IBM lost as much money in 1995 as the rest of the computing industry [b]made[/b].

        The world changes much, much more than it stays the same. Gerstner remade IBM (or, more properly, allowed its more visionary middle- and lower-rank managers remake it) into a dramatically more nimble company than it was then; it's the one large company in the west that's capable of really reaping benefits from the f/oss movement's dynamics.
  • The Point... Is This Good for the Industry?

    I say yes.

    I think this agreement between Microsoft and Novell will be a boon for just about everyone - purchasers, resellers, IT administrators, developers, and end users.

    I'm an IT professional, and think it's heartening to see these two companies not only realizing that virtualization is the future of enterprise
    computing, but embracing it. And by removing patent liability and promising cross-platform support, they have not only opened the floodgates for developers, but will hopefully spawn Linux-based clients from some of the
    leading application developers (e.g., Adobe).
    • re: The Point... Is This Good for the Industry?

      I agree with you, peter_toby. In addition to embracing virtualization, I think that this agreement is a sign that the software industry is realising that web sevices will rapidly disrupt the current software distribution model.

      Innovation and value are increasingly derived from networking and interoperability.

      This shift will benefit all parties and it is reassuring to see that companies are moving towards a coopetition model similar to the one embraced by the semiconductor industry for which IP reuse brought enhanced productivity and faster time to market.
    • I agree ...

      ... it is good for the industry. But corporate relationships are fluid -- and are not based upon friendship. Rather, they are based upon mutual benefit. If either party see that the partnership hurts them more than it helps them, the relationship could be desolved on a moment's notice. Unlike the (almost forgotten) MS-Sun partnership, this is not a partnership of equals. At least not in terms of their viability in the marketplace. Novell is just as likely to be bought up by Microsoft as they are to co-exist with them.
      M Wagner