By Doc Searls
This is an ideal time to start selling Linux "solutions" to enterprises. The Big-Time IT Department allergy to the world's least businesslike operating system has started to wear off, mostly because, as always, the engineers are having their way. They're putting Linux to use all over the place. Management in most cases has little choice but to acquiesce to the obvious, or to ratify it by declaring Linux "strategic."
IBM's decision to enthusiastically adopt Linux is nothing less than huge. Big Blue may not carry the clout it once did, but there still isn't a more significant company in the business. But let's face it: selling Linux to business for IBM hardly requires a sex change. They're a hardware and services company. Linux makes it easier to develop and sell both.
It's a little different for companies that are native to the GNU/Linux community, which grew out of the free software and open source movements. While it would be wrong to call either movement "anti" commercial, it would be fair to say that suiting up Linux for business is not their central concern.
But bringing Linux to business has always been the central concern of Caldera, Red Hat, VA Linux and other Linux community natives. It's also a concern of new companies like Eazel, Zelerate and Ximian. Also of newly arrived computer industry veterans like Borland, HP, Dell and Sun. The difference, again, is in respect for what the Linux community brings to business, whether either side likes it or not.
In every case, what it brings is infrastructure. Linux is now a basic building material for business. So are open source applications like PHP, Squid, Jabber and MySQL, plus open protocols like SOAP, XML and SLP.
There are several fundamental questions facing every company using (or, to use business lingo, leveraging) these infrastructural goods:
It's a given that the native Linux companies are going to want to contribute as much as possible to common code bases. Caldera, for example, is contributing its work on the SLP protocol to the community through OpenSLP.org. Popularizing SLP and inviting more development of the protocol can only help drive sales of Volution, the new Caldera management product that depends on the protocol. It will also help make a market for competing products, which will also raise sales of Volution. Borland is following a similar strategy by open-sourcing CLX, the cross-platform libraries for Kylix, the company's new IDE (integrated development environment).
What's a surprise these days is how many non-Linux (and even non-software) companies are starting to give away infrastructural code. Several months ago Dresdner Bank of Germany gave away the code to a product called OpenAdaptor, which it cost the bank an estimated $5 million to develop. But they believe that the code will do more good for everybody — including Dresdner, which has "first mover" advantage — if it's out there in the open world. The lesson here is that businesses are learning how join the traditional hacker community as contributors to common infrastructure. The more this happens, the more the whole business community begins to understand that what's best for all of us is also what's best for each of us.
But let's go back to native Linux companies like Caldera and Red Hat. These guys are quietly living with the problem of answering #4: what code do we keep closed? Caldera's Volution is a closed-source product. Caldera says Volution is "built on open standards," but that's a buzzphrase that says one thing while avoiding another. The quiet fact is that the source code to Volution is closed. Could Caldera sell it for the same price if it were open? We don't know. Nobody seems to want to test the theory that high-ticket software can also be open source. Red Hat is following a similar strategy with Red Hat Network Services, which quietly deploy Novell's NDS as its "next generation network based storage system." This is no doubt part of the Novell-Red Hat "alliance" that Novell has been eager to celebrate pretty much alone (judging from the comparative sizes of the two companies' stacks of press releases on the matter).
Let's be fair here: this is a touchy area. Open source vendors don't want to risk annoying their development communities. They're moving carefully not only into new product spaces, but into new diplomatic territories as well. (Unless, like Zelerate, Ximian and Eazel, their primary business models are to give away open source software and sell services.)
This will pass as more companies discover the economic paybacks of generous responses to Question #1. But each company will test that question in its own way. IBM has eagerly promised to contribute all it can to the development of Linux and the support of its surrounding open source communities. Meanwhile, Michael Dell has expressed a more conservative position, but one no less carefully considered. Here's what he said during the Q&A after his keynote at the Linux World Expo last August: " Are we going to contribute to the kernel? No, we're not. We're going to use Linux, we are going to provide Linux. As a product becomes established, high volumes require a different kind of company. We have a global support structure. We are spending more R&D dollars, in percentage, on Linux than any other platform."
So there it is. Linux matters. A lot. In fact, it matters so much that it deserves the description that Sun CEO Scott McNealy once gave to the Net: "insurmountable opportunity."