IBM: Boon Or Bane For Linux?

IBM is arguably the most active supporter of Linux: at war with SCO, aggressively contributing to the Linux code base with its own intellectual property, and providing (for them) unprecedented executive-level support for a product the company doesn't actually own. On the other hand, the SCO issue, which wouldn't even exist if it weren't for IBM, has created the largest FUD cloud around the Linux platform.

With IBM's recent divestiture of its PC division, things are about to get really interesting--and the question in my mind is, will the result be positive or negative for Linux?

IBM: The Linux Boon

IBM's PC division was the largest advocate for Microsoft and Intel within the company. A relatively visible and powerful unit, it refused to use the PowerPC architecture and was being dragged kicking and screaming towards Linux. This was largely due to its past involvement with IBM's processor and alternative OS efforts, all of which had been disastrous.

With the divestiture of its PC division, which is expected eventually to pull the XServe line, IBM will be more focused on PowerPC and Linux than ever before. This focus is very similar to the way IBM worked prior to the PC era, when the company owned its platforms and dominated the PC business.

This shift will allow IBM to focus its resources on both platforms and, potentially, to become a true competitor to both Intel and Microsoft. To put this into perspective, realize that IBM Microelectronics has actually captured the entire console gaming microprocessor market, with Sony, Nintendo, and even Microsoft switching to its processor. IBM Microelectronics is also the sole supplier to Apple Computer, the largest desktop Microsoft alternative, and its Linux development efforts are more aggressive than Intel's. In addition, recent breakthroughs have allowed IBM to address more effectively the shortcomings surrounding the Intel architecture, and it is broadly cross licensed with virtually all of Intel's competitors. As a result, in my view, IBM Microelectronics, at least on paper, is Intel's greatest strategic threat.

On the software side, IBM Software has always been the only potential near term nightmare for Microsoft. Until fairly recently, IBM Software was actually larger both in terms of staff and revenue than Microsoft, and it has a breadth similar to the Redmond giant. In addition, if the future of software lies in centralized services (which is very likely), IBM is actually more entrenched than Microsoft. I helped draft the competitive report that was used to justify the spin out of IBM Software in the 1990s (which of course never happened), and the massive amount of research we did at the time indicated that without IBM hardware (which right now is mostly tied to the PC division), IBM Software had the ability to take the market back from Microsoft.

Add IBM Global Services, the most powerful services unit in the world, and suddenly you have the potential of a huge company, focused on Linux with a unique and powerful portfolio of software and hardware offerings.

In effect, you have a company that is close to the relative capability of IBM when it was the Computer industry.

IBM: The Bane of Linux

OK, now step back for a moment and grok what I just said. What is it you like about Linux? The collaborative nature of the product, the idea that is it created by a community, the freedom, the low price? Now think of the technology world the way it was in the 1970s. Was there ever a world farther away from these core values?

If IBM is successful, it will effectively own Linux. It will have the best tools, it will make the most measured contributions to the installed base, and it will bend every rule to make sure its solutions are better, faster, and stronger than anyone else's.

Look at the SCO problem: Didn't it result from IBM's unilateral actions to introduce code into Linux that was at risk? Now take that kind of practice and multiply it a thousand fold. Rather than Linux becoming the new UNIX, doesn't it now become the new AIX, the most non-standard of the UNIX variants?

I've often thought that the Linux community in its blind rush to displace Microsoft was making the same mistake that Microsoft made in its efforts to displace IBM. You often have to become what it is you are fighting, and just as often that means moving into a world you actually don't want to live in.

What made Linux attractive was its freedom in terms of both cost and behavior, the community cooperation, the camaraderie, and, frankly, that it represented a kind of rebel counterculture. But as Linux moves into the mainstream, it ceases to be a rebel and starts to look more like the establishment. If IBM is successful, it becomes the new Microsoft, and Linux becomes the new Windows. In a decade, the emerging generation will attack it the same way the emerging generation today attacks Windows--we will not only have gone full circle, we will have gone around twice.

If you have to become what you hate in order to win, isn't that self-defeating? If the goal of beating Microsoft requires you destroy what is attractive about Linux, does that really make sense? Yet IBM and Linux are on a path which could, in fact, allow the company not only to pummel Microsoft and Windows but to replace them.

Boon or Bane?

How do you define success? If success means defeating Microsoft and Intel by displacing their products with Linux and the Power architecture, then IBM is a huge boon to the platform. If you define success as being able to continue to enjoy the advantages of Linux you have come to know and love, then this move will probably eliminate most of them as IBM takes its greatest weapon, focus, and directs it against Microsoft. Strangely enough, I actually think the biggest beneficiaries of this process will be Microsoft and Intel, both of which lost their focus due to a lack of real competition.

Perhaps it comes back to being careful what you wish for: If, as a child, you want the authority of a parent, you often lose the fun of being a child in the process. As a perennial child myself, I'm not really sure that's a worthwhile tradeoff.