OpenStack by the Numbers: Welcome Kilo!
As April 30th arrived, so did the next named release of OpenStack. Kilo is the 11th official release of OpenStack since its inception in 2010. Born of the with great hopes and minds at Rackspace and NASA, the widely known open cloud ecosystem continues to gain steam, awareness, and sometimes a little bit of criticism. Whether you’re using OpenStack today or investigating it for down the road, this is a good time to take a quick look at what went into the Kilo release.
By the People, For the People
Not even taking into account the corporate support and customer side of the equation, I always like to see how the developer and operations side of OpenStack fares as each release comes out. One thing that is undeniable is that the velocity continues to increase for OpenStack development, and the stability is also doing the same. The program list (aka project list…and probably both because that keeps changing) is widening with the official integration of Ironic, a bare metal provisioning project.
Features are growing rapidly also, with a lot of fixes and enhancements that have been wrapped into this release.
Juno in Review
How does 18,992 code commits sound? That’s pretty impressive if you ask me, and what is just as impressive is if you look at how we assessed Juno in the past here.
So, with Juno in the bank, how about we look at how Kilo measured up by these numbers.
Kilo Kicking it up a Notch
As we took a look at the analytics from activity.openstack.org, the numbers were both impressive and telling. 21,125 code commits from 1,593 developers and a total of 1,839 submitters which covers code, documentation, and training.
Looking at Juno in comparison shows that we have a continued rise above a linear scale with the Kilo release. This is telling as we can see the velocity of development and the growth of the ecosystem which is increasing at each release.
If we look at the picture since the beginning, from 124,499 total code commits, 3,279 total developers, and 3,611 total submitters, it tells us that Kilo produces 16.96% of the total code in just the Kilo release alone!
Code counting doesn’t tell you success stories. What does show a positive story in these numbers is that they keep moving up and to the right. Increasing both on participation, fixes, features, and total growth.
OpenStack Kilo Presentation
As the launch press makes its way around, we will see a lot of good detail on features, advancements, and also on challenges. One thing I do appreciate about a lot of the OpenStack community is that they acknowledge that the challenges are real, and are being taken on as much as possible.
Here is a quick view of the Kilo release as presented by the OpenStack Foundation:
Eric’s Favorite Features
There are lots of very interesting features and improvements that were packed into the Kilo development cycle. According to the OpenStack Foundation there were 394 to be exact. I don’t have an exhaustive list, but I can tell you a couple of things that I am particularly keen on that arrived with Kilo.
Identity federation was available in previous releases, but was well known to be challenging to work with and very new. The Kilo release of Keystone has seen the federation features listed as stable with much more documentation and working examples of how to federate your OpenStack clouds to one another. As people explored OpenStack without federation, the issue rose that if more than one OpenStack cloud was implemented that the identity environments were separated, and thus difficult to manage.
By adding stable Keystone to Keystone federation we have the ability to create a true hybrid OpenStack cloud. This is also important as companies are embracing OpenStack and the potential for two organizations to have to merge during a purchase or partnership was a bit limiter up to this point.
Ironic for VirtualBox (experimental)
Although this is an experimental feature within the Ironic (bare-metal provisioning) program, the availability of a way to test out Ironic deployment for people is going to be very helpful as we get more folks exploring OpenStack. It’s one thing to have a working OpenStack environment, and a whole other thing to have procedures and methods to build, deploy, and grow the infrastructure. Using VirtualBox as the test bed for Ironic will allow more people to test the waters and provide feedback to the development teams. This can only lead to good things.
Scale Preparation for Neutron
This is a work in progress, but it is top of mind as we head into the Liberty release. Networking within the OpenStack environment has often been shown to have scale challenges when growing in large environments. Many won’t hit any of the scale limits that are currently a challenge, but it is important just the same that the teams are working to address this and have done lots of work to build in future improvements as upcoming releases come out.
Yes, people are a feature. This includes a growing community, an eager developer ecosystem, and the most important thing of all which is people using OpenStack. The focus on the operator and what is called the “superuser” has been big with the last couple of OpenStack Summits, and with good reason. The more people who are using OpenStack in some form, the better the overall ecosystem will do.
Looking forward to an equally exciting Liberty release, and I hope to see lots of you in Vancouver at the OpenStack Summit!