[openstack-dev] [all] The future of the integrated release

Stefano Maffulli stefano at openstack.org
Fri Aug 22 00:23:10 UTC 2014

I think we can't throw Ceilometer and Triple-O in the same discussion:
they're two separate issues IMHO, with different root causes and
therefore different solutions.

On 08/21/2014 06:27 AM, Jay Pipes wrote:
> The point I've been making is
> that by the TC continuing to bless only the Ceilometer project as the
> OpenStack Way of Metering, I think we do a disservice to our users by
> picking a winner in a space that is clearly still unsettled.

When Ceilometer started there was nothing in that area. A quite
significant team formed to discuss API and implementation for metering
OpenStack. All was done in the public and getting as many people
involved from the start. Ceilometer was integrated because the whole
OpenStack project has been about fostering collaboration *on top* of a
common technical infrastructure. And because it was considered ready
from the technical aspect.

Now we're finding out that we don't have appropriate processes and tools
to evaluate what happens later in the maturating cycle of technology:
Ceilometer is not considered the "only" nor "the best" tool in town
anymore. How to deal with this?

Your proposal seem to go towards the old, known, 20 years old territory:
OpenStack should provides infrastructure and some mentorship to
development teams, and that's it. That's somewhere between SourceForge
and Apache Foundation.

Contrary to other open source foundations, we have put in places
processes and tools to pick our favorite projects based not just on
technical merits. We're giving strong incentives to do open
collaboration. The collaboration across large corporations as put in
place in OpenStack doesn't happen by chance, quite the contrary.

This is what makes OpenStack different (and probably one of its main
reasons for success).

> Specifically for Triple-O, by making the Deployment program == Triple-O,
> the TC has picked the disk-image-based deployment of an undercloud
> design as The OpenStack Way of Deployment. 

Triple-O is a different case, as puppet and chef modules have existed
since I can remember and OOO is only one of many. Different problem that
should be discussed at TC level. A lot of improvements can be imagined
for the Deployment program.

> I believe by picking winners in unsettled spaces, we add more to the
> confusion of users than having >1 option for doing something.

I think this conversation is crossing the very important "What is
OpenStack?" issue. Most likely the answer won't be as clear cut as
"OpenStack is what the TC decides is integrated". A lot of your concerns
about the TC 'picking the winner' will be resolved because it won't be
the TC alone to decide what will be using the OpenStack mark.


Ask and answer questions on https://ask.openstack.org

More information about the OpenStack-dev mailing list