CentOS Named to be a Red Hat Sponsored Project

Holy crap, this is big news, for a number of reasons. I’ll give you my take on it below, but I’ll also add links from folks that I trust in order to provide additional takes on this momentous occasion.

I think this is a great move for both Red Hat and CentOS, provided everyone understands the target audience and target platforms for each distribution. And by everyone, I mean everyone. Internal folks at Red Hat (sales and techies), customers /users of RHEL, and users of Fedora and/or CentOS all need to understand that little is changing for the distributions, but a great deal of tech enablement will be the result.


In prior years, CentOS was simultaneously a thorn and a gift to Red Hat. Some customers would ask why they should pay for RHEL when CentOS was free and compatible with their applications. (Anyone concerned with security and timely patches were quickly re-convinced, along with a multitude of other Red Hat values.) Identical configuration and administration coupled with near-identical kernels meant most applications worked fine. Others started with CentOS when they had zero budget and then migrated to RHEL as they started making money. This was certainly the case when I worked at Red Hat in professional services.

But to be truthful, Red Hat was also it’s own thorn, more so in the more recent years. Want to play with OpenShift, Gluster, RDO, or oVirt? These are all phenomenal community supported and developed projects that didn’t have a natural Red Hat-based platform to deploy these community projects on. Fedora is great for bleeding edge and can even serve as a production platform. And RHEL is certainly capable, but why are you going to deploy a free app on a paid Linux distro?

For example, what if you want to deploy RDO on a stable platform? RHEL is not an obvious choice if paying for support subscriptions isn’t in your budget. And Fedora is not a natural choice either as it’s package versions may be far beyond what RDO (or gluster, or oVirt, etc) supports. Reconciling old and new packages in order to get an app to work shouldn’t be part of the equation, especially if you’re just venturing into the waters. Let’s lower the bar to entry! Let’s lower the bar to continued involvement!

(end sidebar)

My take, understanding, and point-of-view is as follows:

Fedora remains the community supported upstream to RHEL. Red Hat will continue to develop & harden features in Fedora for potential release in RHEL. Users and developers can continue to rely on Fedora to be the “bleeding edge” of the RHEL-based distributions. This does not change. Fedora is here to stay.

RHEL remains the Red Hat supported platform for enterprise, commercial, public sector, and all other customers requiring full support for their platform and applications. Need   support for that tier 1 database, a stable development platform for you application servers, or just minimal support for your environment for when things go bump in the night? This does not change. RHEL remains the fully supported Linux distribution of choice in the enterprise.

Since 2004, CentOS was the de facto standard for deploying applications on a free RHEL-based platform. This too remains the same, if not reinforced. CentOS is now poised to become a “parallel downstream” to RHEL (if I may coin a phrase) in that it is built from RHEL, but in a manner that is compatible with RHEL in order to better serve other community supported projects such as oVirt, RDO, Gluster, etc. It also means that a customer that wants to move their CentOS-based app stack from sandbox to production, they can do so with minimal interruption.

This really does lower the bar to entry and lower the bar to continued involvement. If nothing else, it also allows you to not think about the platform, and focus more on the application or application development.  So whether you are self-supporting, or just getting started with the technologies, Red Hat pulling in CentOS as a full partner and sponsored project will make things that much easier when it comes to deployment and ongoing maintenance. This has certainly changed how CentOS can be used.

Want to hear some other perspectives? Here are a few:

Computer Weekly has a short article.

Robyn Bergeron (Fedora Community lead and friend) has a great take on things in her blog, including links to a FAQ on the subject

And my close friend Richard Morrell (Cloud Evangelist) has a great podcast on this, including a chat with Karanbir Singh, the CentOS project lead

hope this helps,

Captain KVM


Agree? Disagree? Something to add to the conversation?