In February 1999, IBM announced it would support Linux and a partnership with Red Hat. By July of that year, Advanced Communication Design, a developer and OEM of in-store interactive digital audio and video merchandising systems, made history by being one of the first companies to deploy a major, mission critical IBM/Red Hat Linux system. Why? Marco Scibora, ACD’s president explained then it was because “Linux is very reliable, and its help resources are extremely fast,” and also, “now that Linux has major corporate help support from IBM and other companies, it makes a great environment for customized programs.”
Fast-forward to October 2000: IBM announced it had a grand operating system unification plan for its servers. Its name was Linux, and with the partnership of the four major business Linux distributors, Caldera, Red Hat, SuSE and Turbolinux, IBM made it happen. Today, there’s no modern IBM hardware from laptop to mainframe that you can’t run Linux on.
Now it all seems inevitable. IBM and Linux, Linux and IBM. Today, the two go together as closely as Microsoft and Windows. But what is IBM really getting from Linux? How well is the partnership between what was once seen as the stodgiest of all computer companies and the most rebellious of all operating systems actually going beyond the ad campaigns and the constant announcements of new deals?
Those famous billion dollars
The example people usually cite to such a question is IBM’s famous $1 billion spent in 2001 on Linux. The figure is certainly an impressive one and even more striking, is that Bill Zeitler, IBM’s senior vice president and group executive for eServer, claims: “We’ve recouped most of it in the first year in sales of software and systems.”
What Zeitler didn’t say, though, was the billion-dollar figure isn’t just Linux, as both Stacy Quandt, Giga Information Group’s Open Source analyst, and Dan Kusnetzky, IDC’s vice president of system software research, point out. It’s for the division that also works on AIX and Caldera’s OpenUnix. While the OpenUnix segment is tiny, AIX, the operating system of choice for the pSeries (aka, RS/6000, POWER systems) is a very popular mid-range and clustering Unix choice and competes head to head with Sun’s Solaris and HP’s HP/UX.
So what is the actual breakdown? IBM’s not telling. According to an IBM senior PR representative, “The $1 billion investment IBM made in Linux in 2001 was across hardware, software and services. The investment spanned a number of areas, including development, marketing, sales, ISV support and advertising worldwide. As for a breakdown of the investment between AIX and Linux, we don’t break the percentages down to that level.”
You’d think IBM must know the numbers, but one is led to think that while Linux doubtlessly gets hundreds of millions in support, perhaps even the majority of that billion, that impressive figure doesn’t tell the whole story.
No one doubts Linux’s importance to IBM. Even Simon Phipps, Sun’s chief technology evangelist, says, “IBM is selling a lot of Linux.” He also says, “They also have things going on out there that aren’t entirely kosher.”
For a specific example, Shahin Khan, Sun’s chief competitive officer writes: “The economics just don’t work. IBM claims it is financially justifiable to consolidate as few as 20 Linux servers on a z800 (a new zSeries mainframe). With an estimated starting price of $400,000 for a z800 with a single CPU engine enabled, that claim seems exaggerated compared to Linux servers that hover in the $1,000 to $2,000 range … When customers realize Linux on mainframe utilization will be low, and administration costs have still not been factored in, you can begin to see how the costs will add up. And let’s not forget the support costs that will need to be purchased, either from the distributor or IBM Global Services.”
Of course, Sun has its own ax to grind here. Still, Quandt observes that IBM hasn’t been forthcoming with where its almost billion dollar return from Linux has been coming in, either. She speculates, “Even a few mainframe sales could go a long way towards explaining that number.”
Has IBM been good for Linux development?
Here there can be little doubt. It has supported porting Linux to its own platforms and development of device drivers; has made its journaled file system technology available under GPL; and has supported Linux developers. It’s easy to see why many analysts and developers think that IBM is the best friend a Linux programmer can have. And, that’s before the money and resources IBM provides to its Linux distribution partners.
But, as Bruce Perens, HP’s senior strategist for Linux and Open Source, points out, it’s not all sweetness and light for IBM, Open Source, and Linux. He thinks that, “IBM has one huge skeleton in its closet. Its strategic intellectual property policy is anti-Open Source. IBM’s IP license structure is a fiefdom not responsible to another and the licensing department has a lot to say. Where does IBM Open Source and Linux communities and IBM’s IP policy groups meet? To me, they seem to only meet at IBM’s CEO.”
What Perens fears is that IBM may veer away from Open Source orthodoxy. While he works for an IBM competitor, he does have a point. In areas other than Linux, IBM has recently had a checkered history with Open Source and open standards. For example, IBM and Sun have recently been fighting over Sun not being allowed to join as a senior member of the open Web Standards ubergroup, the Web Services Interoperability Organization and over their now slowly settling fight over open source Java tool development kits. On one side, Sun was promoting NetBeans, while on the other, IBM was aggressively moving forward with its new Eclipse platform. They had the same goal, both Open Source, but there was a short, nasty political fight, which didn’t bring anyone closer to usable code.
Will IBM back off from its currently Open Source supportive Linux stance? So far, it doesn’t look that way. Its fights with Sun probably, many observers think, have more to do with its long running competition with Sun than being able to promote a closed source, closed standard policy.
Even Perens is quick to point out that IBM and HP’s adoption of the GPL has lead to the two becoming “good collaborators.” He adds: “The GPL is the only thing that ensures that partners can work together. Traditionally, HP and IBM haven’t worked well together, but in Open Source, we manage to do it.” Between Open Source making it possible for IBM to work with other industry powers and empowering its Linux operation, it’s hard to see IBM going back to its old proprietary ways.
Business wise, the case is open and shut. IBM has proven to be a winner for Linux. As Eric Raymond, co-founder of the Open Source Initiative, says, “Having the biggest technology company in the world back Linux sent a clear signal to a lot of conservative types at large corporations who would otherwise have taken longer to pay serious attention.”
Perens agrees, “I like IBM’s advertising because it gets me in to see senators and congress people. Before, they had no idea what Linux was or that it was important. Now, they do.”
Kusnetzky explains the power of IBM in detail. “When most enterprises look for a new system, the first thing they do is to decide on databases and tools. The next decision point is what applications will work with your DBMS and tools for those things we don’t want to build for ourselves. The hardware and operating system are in their third round. So, for example if Oracle is the database, then products that don’t work well with Oracle don’t even show up. Others start with packaged applications to minimize expenses, and then go to the database, and only then get vendors in the room, and decide on an operating system and architecture.”
He goes on, “That’s why Linux has had trouble entering into enterprise space. Quite often the Linux suppliers don’t have the relationship with independent software vendors. So when the party is thrown they don’t always get an invitation. Sometimes, Linux may be better choice, but it may never get a chance. Linux people often don’t have marketing folks, if the decision maker is an engineer or tech type, then Linux’s technology will win out for them. However, talking technology to the CIO isn’t going to work. You have to talk to Dilbert’s boss, not Dilbert. What IBM has done is do a better job of convincing both Dilbert and Dilbert’s boss that Linux is something they should consider.”
And, with IBM, Dilbert, and Dilbert’s boss, all interested in Linux, Linux finally is showing up in places beyond Web servers and departmental file/print servers. But where exactly are IBM, Linux and partners going?