Is 10 years of RHEL support still sufficient ?

Submitted by dag on Fri, 2010/08/20 - 08:31

Almost exactly a year ago, I posted a blog article titled Is 7 years of RHEL support still sufficient ?. In that article I make the case that with RHEL major releases moving from 1.5 years to 3 years and virtualization reducing the importance of hardware life cycles, RHEL support should be extended beyond 7 years.

Yesterday Red Hat announced that it did just that. From today Red Hat offers RHEL Extended Life Cycle Support (ELS) on top of normal subscriptions for specific versions and variants. Meaning that for RHEL3 only i686 AS/ES and for RHEL4 i686 and x86_64 AS/ES and ia64 AS are taken into consideration and are thus more expensive than regular support. No pricing information is available yet.

As a result of this the timeline of RHEL releases looks like this:

Click the picture to see it in all its glory

Honestly, the announcement comes too late for a lot of companies that already have planned and executed (sometimes painful) RHEL3 migrations. Extended support could in some of these cases be a more cost-effective solution. On the other hand, with this late announcement Red Hat might have anticipated lowering the number of applicants at the start of the offering. And a sufficient high price tag will influence decision makers even more...

An offering like this also risks that application vendors might hold on to (too) old RHEL releases rather than spending time testing and fixing their software (often agents) on the latest RHEL release. Which is one of the most important reasons why companies delay implementing newer RHEL releases today. Which makes Open Source alternatives more attractive (as they do not suffer from that). Companies that suffer from this delayed availability of proprietary software should complain loudly to vendors and threaten to move (or simply move) to alternatives.

All in all, this new support option clearly strengthens Red Hat's position in the wider Enterprise OS market and makes their virtualization product (RHEV) offering even more attractive compared to the competition.

And to answer the question, yes, 10 years of support should suffice for some time ;-)

like windows

In a perfect world, we would plan the upgrade path/end of life during initial system setup. So, the annoucement is way too late.

But at least, RHEL now has (optionally) the same life cycle as Win2k8; see also http://benjamin-schweizer.de/operating-systems-lifecycle-chart.html

Better life cycle model

I'm hitting the same problem as you do and lots of RHEL users do.

Red Hat should switch to a life cycle model like Next-release +5years. With such a model one does not care if RHEL6 appears in 2010 or 2011.

As you mentioned correctly, the time between RHEL releases slowed down from 18 months to almost three years. Hopefully this slow down will at least stop, otherwise RHEL7 will appear in ~2014

Again: Red Hat needs to switch to the model "Next-release +5years" like I wrote in an article ealier this yes: http://blog.delouw.ch/2010/02/28/where-the-heck-is-rhel6/

I totally agree with

I totally agree with you.

Whatever the support duration, if you want to always upgrade from RHEL-n to RHEL-n+1, you still have to upgrade your servers every 3 years (on average, in the long term). Which is impossible^M difficult for a regular company.

Cyrille