This week has been great engaging with OpenStack experts from a wide range of industries and disciplines. There are two major themes that became really clear about OpenStack this week:
- The growth and user adoption is huge. Representatives from Bloomberg, BestBuy, Comcast, and others, spoke at one of the keynotes about the many ways in which they are using OpenStack in production today. This not only shows great growth, but the cases of real world use validate it. Unlike many open source projects that start with great momentum and fall off, OpenStack isn’t one of those. The corporate support and many code contributors continue to advance and keep this project alive and well.
- Direction is unclear, yet very clear. I like to think of OpenStack very much like Linux; open source in nature and made up of a handful of “must have pieces.” Yet there are many distributions of Linux, not one end-all-be-all distribution. Like Red Hat has commercialized Linux with the support model and their RHEL, I can see the same thing happening with OpenStack. Many vendors are playing in OpenStack and putting their secret sauce on it to bundle a service or solution.
One of the sessions that was very entertaining and painful to watch (like a train wreck, you just have to look) was the Ops Panel: Reference Architecture. It was a fantastic session full of opinions and heated debates. Myself coming from the Ops space, I too had my own opinions and found myself going back and forth, in my head, over whose side I was on. They are designed to be Reference, yet they are packaged as turn-key solutions, almost to the point of driving to claim a de facto standard. The panel was rich with Cisco, HP, Dell and, my personal favorite panelist, Randy Bias, Co-founder and CTO, Cloudscaling. Unlike the other panelists, Randy comes from the space of no vendor lock-in and keeping everything open. So, he had a unique perspective, not coming from a vendor selling hardware.
This session was just one of many that I attended that exposed the rapid adoption, yet unclear and clear direction of where and how OpenStack will end up. OpenStack is making changes and releasing new features, which seems to happen almost every day. I’m of the camp that the final offering which makes up OpenStack from Nova, Keystone, Glance, Swift, Horizon, Quantum, and Cinder needs to get closer together and drive much closer to a common goal. A good example of this separation is the Nova networking and Quantum. What is going to happen there? Does Nova networking go away and Quantum fill in the gaps, or will there be a place for them both in the ecosystem? Then you have pieces like Heat and Ceilometer: how will or won’t they merge into Horizon? Still lots of questions that are yet to be answered.
For those who would like OpenStack lexicon defined, here’s a post that does a nice job: http://ken.pepple.info/