Debian Linux to Get Down to Business?
by Timothy Prickett Morgan
While lots of people in the business community talk about Linux these days, and while the various Linux distributions that are based on the Debian offshoot of Linux are probably second only to Red Hat in terms of installed base, you just don't hear businesses talking about Debian. But, if the people behind Debian have anything to do about it, Debian is going to get its act together and quite possibly become a real contender in the Linux racket.
I know, I know. You've heard this before, and so have I. In August 2004, in fact, I wrote a story about how Bruce Perens, the former project leader for the Debian Linux distribution, had announced UserLinux, which aimed to take on Red Hat and Novell in the enterprise Linux market. At the time, Perens said that while commercial Linux distributions were cheaper than Unix or Windows alternatives, they were nonetheless far too expensive and that high prices were, in fact, limiting the uptake of Linux among commercial customers. It is almost a year since UserLinux, which is a streamlined version of Debian, made its splash and almost two years since Perens put out the idea. And according to the UserLinux site, UserLinux Enterprise Server, Enterprise Desktop, GUI Server, and SOHO Desktop--the four variants of UserLinux--are still in beta. And no one talks much about UserLinux any more.
That includes Ian Murdock, who is the "ian" part of Deb-ian, the founder of the Debian distribution, and the current chairman of commercial Linux distributor Progeny. But what Murdock is talking about is getting the Debian collective in order and working towards the goal of creating a single Debian core that all Debian-based distros customize. Murdock started the tongues a-wagging in June when he posted some musings in his blog, "Debian: Where
should wewe should go from here ?". The Debian "sarge" update was released in early June, the first major upgrade of Debian in years. And in that blog posting, Murdock argued that what Debian really needed was two things: to get a timed release cycle in place like all professional software development projects have and to get Debian united around a common core.
The Linux community has seen this movie twice already. First with the UnitedLinux effort put together by the then-independent SUSE (now part of Novell , the then-independent Conectiva (just eaten by Mandriva , the then-Linux distro Caldera (now known as the litigious Unix vendor SCO Group), and the still independent but largely sidelined Turbolinux. These four decided to base their Linux distributions largely on the SUSE Linux core, and then all hell broke loose after the SCO suit and after Novell bought SUSE, and UnitedLinux died a quiet death. The second time we have seen this movie is the Linux Core Consortium, which was announced last November by Conectiva, Mandrakesoft, Turbolinux, and Progeny to implement the Linux Standards Base 2.0 specification jointly as a shared set of source and binaries that they would then all distribute as the heart of their own Linux distributions.
The fact that Murdock is now talking about creating something called the Debian Core Consortium and is emphasizing the benefits of Debian as implemented as Componentized Linux pretty much tells you that for all intents and purposes, the LCC is kaput and Murdock wants to corral all the of the Debian distros into using Componentized Linux, which is billed as an improved implementation of Debian that allows mass customization. It is not a coincidence that the City of Munich in Germany has created its LiMux variant on Debian, and nor is it a coincidence that the city of Extremadura in Spain has created its own LinEx variant on Debian, either. Murdock is on to something, and that something is called mass customization.
Componentized Linux was an internal way that Progeny built customized Linuxes for its customers, who sometimes want appliance servers, infrastructure servers, desktops, or embedded systems. Murdock simultaneously steered Debian sarge out the door and helped a team of coders productize that customization methodology to create Componentized Linux, or CL for short. And now, what Murdock seems to want to do is form a consortium of Debian distros who adopt CL as their core, use its customization features to make specific Debian distros for specific market segments and customers, and all pull the rope in the same direction. Murdock's plan was to announce the Debian Core Consortium at LinuxWorld in San Francisco in a few weeks, but this being the open source community, there are few secrets--or at least few that stay that way for very long.
For now, Progeny is working on taking the Debian sarge release and turning it into the heart of Componentized Linux. The plan calls for CL 3.0 Pre Release 3 (PR3) to ship today--PR1 shipped on April 28 and PR2 shipped on June 2. CL 3.0 Release Candidate 1 (RC1) is expected to come out on August 2, and CL 3 Release 1.0 is expected to debut on August 9. Basically, this software is already cooked. CL 3 Release 1 will be the heart of Progeny Debian, and it will be certified to the just-released Linux Standards Base 3.0 specification. The Debian Core Consortium is by no means walking away from the Linux Standards Base; if anything, it is becoming the standard bearer for the standard (pun intended). To get a detailed description of Componentized Linux, see http://componentizedlinux.org/roadmap/.
It will be interesting to see how the Debian distributions react to Murdock's proposal for a Debian Core Consortium. But interestingly, maybe this time, the distros won't matter as much as the companies and the governments who, for whatever reason, do not want Red Hat, Novell, Mandriva, or Turbolinux distributions and who decide they want to make their own customized distribution. Componentized Linux and the Debian Core Consortium will give them the means by which they can do this and still maintain compatibility with Debian.
UserLinux: A Cheaper Enterprise Linux?
Linux Core Consortium: Déjà Vu All Over Again