It appears that Microsoft wasn't the only target on IBM's radar screen when it decided to adopt Linux and tie together all of its platforms. A little birdy very close to the company informs me that Microsoft was an afterthought. The real target was Sun.
by Ed Sperling
When you consider what IBM sees as its biggest threat - namely its hardware sales - it's hard to argue otherwise. Sun has done to IBM in the Web-server market what Digital Equipment Corp. did in the minicomputer market - it swooped down and stole IBM's lunch right out from under its nose.
Who was minding the store? Apparently, no one. IBM almost single-handedly made e-business a household word, and Sun made billions off the concept. Not a bad haul for a firm that was running out of steam as a work station company. IBM started out adding Linux hooks into its middleware - mostly MQSeries - as part of a skunk works project. The company wasn't even sure it would work. But it did, and someone there got the bright idea that it would open up a huge opportunity for the company's servers.
If ever there was a company with a legacy problem, IBM is it.
Because it basically pioneered the development of mainframes and PCs, and added RISC systems and minicomputers along the way, each OS was developed from scratch on its own, with no thought about making everything work together. There's a reason that IBM is the leader in middleware. It has to be. None of its platforms is compatible, and none of the projects it launched in an effort to make everything work together (Systems Application Architecture, Office vision LAN, and the Computer Desktop Environment) ever made it out the door. But when Linux came knocking, IBM saw a big opportunity.
For one thing, Linux could be made to run on every IBM platform. But equally as important in Big Blue's strategy is the realization that if Linux gains enough momentum, it will derail Sun's position with Solaris and level the playing field in Unix servers. At that point, it's merely a question of choosing your hardware, and IBM can throw a big array of Linux servers at any business problem—everything from a mainframe to a Netfinity PC using standard Intel components.
The Linux plan also solves another of IBM's nagging headaches. Ever since IBM began developing more than one platform, first the company's sales force and then its business partners have had to deal with some serious political battles. AS/400 groups have undercut RS/6000 sales, and vice versa. And the mainframe group has always been sacred, as much for historical reasons as for the profits to be made on big iron.
With 1,700 downloads of Linux for System/390 mainframes so far, IBM's plan seems to be catching fire. Maybe my next tip will come from a penguin.