Is 7 years of RHEL support still sufficient ?

Submitted by dag on Wed, 2009/09/02 - 20:56

Seven years is a long time. Most people won't know where they are in 7 years. I'm interested to find out how many kids will be running around the house in 2016 :-) However if you are responsible for thousands of Enterprise Linux servers, seven years may no longer be sufficient.

Let me explain what I mean. When Red Hat released RHEL2.1, seven years of support was perfect, seemed more than one would want. RHEL3 came 18 months after RHEL2.1 and after one year of testing RHEL3 and 3rd party integration new systems could be deployed, giving you 6 years of support. Your hardware would usually not outlive the operating system support.

Big companies and organizations usually have a hardware cycle of about 4 years, after which the support of the hardware becomes too expensive to maintain. (Think: spare parts) This means that if new hardware arrives, you need to make sure that what you load on these servers will be supported for at least 4 years. If you take virtualization into scope, this cycle may be virtually stretched.

Now look at the following RHEL facts:

  • RHEL2.1 released on May 2002, EOL on May 2009
  • RHEL3 released on Oct 2003, EOL on Oct 2010
  • RHEL4 released on Feb 2005, EOL on Feb 2012
  • RHEL5 released on March 2007, EOL on Mar 2014
  • RHEL6 anticipated for Q1/2010

From this you can see that the releases for each RHEL major version went from 1.5 year, to 2 years, to almost 3 years. On the following chart this is more clear:

Click the picture to see it in all its glory

So today if we need to load a new server, RHEL3 and RHEL4 are out of the question (EOL in less than 2.5 years). The only viable option for installations is RHEL5, still supported until March 2014. Which is slightly more than 4 years today, but when RHEL6 is released it will not be immediately ready for new installations either, it takes 1 year to 18 months for software vendors to start supporting a new operating system.

So what is a system administrator to do in June 2010 if he wants to deploy RHEL or CentOS ? Ship RHEL5 that will be deprecated before the hardware is up for replacement, or ship RHEL6 that may not integrate well, or instead use a (yet) unsupported combination ?

In a vaguely similar situation with RHEL4 going out of "installation support" before RHEL5 was ready for deployment, Red Hat decided to extend this installation support for an additional year to 4 years. With RHEL5 we are facing a problem with the total support time so there is only one thing to do. Extend the total support of RHEL5 and RHEL6 to 9 years instead.

It makes a lot of sense, now that Linux in general is maturing (and more specifically, the kernel is not having major bumpy releases) and new main RHEL releases are slowing down. Less reasons for a new major release. Less effort to back-port functionality and features. In turn, Red Hat can better deploy its existing resources to fewer supported releases, but instead support them longer.

Everyone wins in the end.