Last month I used this space to talk about IBM/Red Hat’s plan to restrict access to Red Hat Enterprise Linux (RHEL) source code. This eerie announcement, which seemed quite contrary to the ideals of free software, sent shock waves through the community.
Last month I used this space to talk about IBM/Red Hat’s plan to restrict access to Red Hat Enterprise Linux (RHEL) source code. This eerie announcement, which seemed quite contrary to the ideals of free software, sent shock waves through the community. Some said it violated the spirit of the GPL, and others argued it was necessary to stop the clones from stealing Red Hat’s business. Everyone agreed that Red Hat had developed a novel argument that could potentially allow them to skirt around the code-sharing protections of the GPL, and the general feeling was that the matter would only be settled after a protracted courtroom battle.
Regardless of where this episode ends legally, it is now clear that Red Hat’s clones and other competitors are not planning to wait for the courts. Various distros have come up with various plans, some of which I covered last month. This month, the big news is that Oracle, SUSE, and CIQ have joined forces to launch the Open Enterprise Linux Association (OpenELA).
OpenELA refers to itself as “a collaborative trade association to encourage the development of distributions compatible with Red Hat Enterprise Linux (RHEL) by providing open and free Enterprise Linux (EL) source code” . It would take a long time to explain why this organization would be able to provide access to Red-Hat-compatible source code when Red Hat itself restricts access. Suffice it to say that Red Hat figured out a legal hack to the GPL, and the companies behind OpenELA have several options for how to hack the hack.
The legal arguments will have to play out in court – I’m more interested in what this new organization is, what it will do, and whether or not it will succeed. OpenELA is exciting for a number of reasons. First of all, it ensures ongoing free access to the Enterprise Linux code base, which will help to avoid the fragmentation and needless incompatibility that often confounds Linux users. Another important benefit of this change is that it reasserts the free software vision just when it seemed to be slipping away. The GPL is supposed to be eternally self-correcting. No vendor can corner the market, because if they try to restrict access, the community responds by forking the code and offering alternatives.
So far so good, but a word of caution: There are many complications to companies teaming up to produce a shared product that is vital to their individual livelihoods. It is way more difficult to maintain a full enterprise Linux distribution than it is to write a check every year to the Apache Software Foundation or send a few developers to work on the kernel. Ultimately, each of the companies participating in OpenELA will have to sublimate their own priorities for the project to stay on track.
Back in 2005, a group of Debian-derivative distros announced that they were banding together to form the Debian Common Core (DCC) Alliance , which would work communally to provide a foundation of common components they hoped would streamline development and “encourage commercial adoption” of Debian-based systems. As soon as they started, though, it became clear why the participants were separate distros in the first place and not a single Linux. The DCC Alliance was fraught with disagreements and only lasted for two years. Admittedly, some of the companies putting money into the project were having their own financial issues (who remembers Xandros and Linspire?) But the fact is, a project of this magnitude requires hundreds of decisions, and there are many reasons why different companies would want to make those decisions in different ways. Companies don’t make money by sharing everything – they make money by differentiating. When corporations try to collaborate and compete at the same time, they sometimes end up playing musical chairs like the generals in Evita.
Oracle and SUSE, for instance, aren’t exactly best bunkmates. It is true that SUSE supports Oracle database systems, but it is also true that Oracle likes to claim “Oracle database runs best on Oracle Linux” . SUSE, on the other hand, is the leading system for supporting SAP’s HANA database and ERP software, which competes directly with Oracle’s Fusion Cloud ERP suite. CIQ is a smaller player than the others, but one of their areas of interest is HPC, which has long been a strength for SUSE.
The vendors behind OpenELA will have to stay together and keep their eyes on the prize if they want to avoid slipping into a game of musical chairs.
Joe Casad, Editor in Chief