Red Hat’s new model

A feature of free and open source software is that anyone can make a copy of the software, rebrand it, market it and sell it to the world...

The centOS community used this facility to build a successful clone of red hat enterprise Linux. Red Hat has now taken CentOS under its wing.

Traditionally software companies have relied upon the ‘unique’ qualities of their software as the selling point, but selling free software is a different proposition. Most, but not all, of the distributions that Distrowatch currently lists began life as copies or derivatives of one or other of the generic Linux distributions – Red Hat, Slackware or Debian – each of which owed some kind of a debt to Linux pre-history in the shape of SLS or Owen le Blanc’s MCC Interim Linux, which is often claimed to be the first installable Linux distribution.

The primary role of a commercial server-based GNU/Linux distribution is not the sale of the software but of follow-up services – subscription, consultancy, installation, training, support, upgrades and maintenance. Advances to free software may come from any number of sources, but for a distribution to succeed it has to have something that differentiates it from the others, an extra layer of polish or a loyal following, and a reputation for providing good service.

A GNU/Linux distribution is a collaborative effort that may include the work of thousands of individual and corporate contributors. To make a success of the follow-up services that sustain an open source software company, the package of software it provides has to be of sufficient quality to justify purchase of subscriptions and services in the first place.

Red Hat owns the brand and the quality assurance that goes with the Red Hat trademark, but does not ‘own’ the software it sells. For this reason, CentOS and Oracle have been able to distribute rebranded copies of Red Hat Enterprise Linux, which provide ‘complete upstream compatibility’ with Red Hat’s product without fear of legal approbation. Red Hat is not selling software, but quality assurance, secure upgrades, training and support, and licensing of third-party software, a model that was pioneered by Cygnus Solutions and Michael Tiemann, who is now vice president of open source affairs at Red Hat.

The pay-off for Red Hat is that the cumulative R&D value of the packages that comprise a Red Hat Linux distribution stretches into billions of dollars, and is the work of a multitude of private individuals, developer communities, corporate interests and other GNU/Linux distributions. What differentiates Red Hat from its competitors is the distinctive way it is put together, and the quality of the service it provides for its customers.

The pay-off for the end user is that the barriers to entry are low. If a GNU/Linux distribution lets you down, you can switch to another version of Linux and hope for some equivalence, so it is incumbent on a commercial GNU/Linux distribution to be at least a match for its competitors, honest and true to its developer communities, and on the ball with the services it provides.

CentOS has been adopted as an alternative to RHEL by a large community of organisations and companies who have the in-house expertise to support a Linux server installation and no requirement for the quality assurance and support provided by Red Hat. CentOS rivals Debian as the most popular server installation of Linux. For this reason, Red Hat’s recent decision to acquire CentOS makes a lot of sense.

Red Hat gains a user base that is already effectively part of its ecosystem, who may be seen as potential clients for Red Hat services into the future, and access to an experienced user community with direct experience of a RHEL clone in the field. Effectively, Red Hat will be providing a ‘free’ version of its enterprise operating system, albeit one that lags some way behind RHEL, in return for feedback from the community. A secondary benefit is that it neutralises the possibility of the purchase of CentOS, and its wider ecosystem, by a competitor.

Red Hat has been a lucky company with a knack of being in the right place at the right time, epitomised by its transition through a successful IPO from being the foremost community-based Linux of the Nineties to the most successful provider of open source server systems to the corporate world during the last decade. Red Hat was able to capitalise on a willing corporate market at a time when Linux was becoming the logical choice for back room servers, but it was able to do this through the good relations it maintained with its user and developer communities.

Some people have seen CentOS as a threat to the Fedora community, but Fedora has a distinctive and unambiguous role in Red Hat’s ecosystem. Fedora’s purpose is to be on the bleeding edge of free software distributions, prototyping ideas and implementations that may or may not be adopted in future editions of Red Hat’s commercial distributions. RHEL is developed from a snapshot of Fedora.

Equally, some have seen Red Hat’s purchase of CentOS as signalling a death knell for CentOS, but this would not serve Red Hat’s purpose well. CentOS is not a competitor for Fedora, and remains a clone of RHEL, differentiated only by its branding. The appeal of CentOS is that it provides Red Hat with a ready-made pool of experienced users of Red Hat software, and for Red Hat’s purposes into the future, the more users it has the better its model works, and the success of CentOS is to be maintained and promoted.