Crucial Strategy for OpenStack in 2017

I have been working on OpenStack full time for 5 years. During that time I have seen a meteoric rise in OpenStack in terms of adoption and feature parity with OpenStack’s largest competitors: AWS and VMWare. I am running for the OpenStack board of directors as an individual candidate to improve the competitive outcomes for OpenStack.

In my technical work I have been involved in the founding of OpenStack Heat, which was an effort to produce parity with AWS CloudFormation, while also blazing a trail on the incubation track for OpenStack. Without Heat and the excellent engineering team which implemented the first incubation project, OpenStack may have taken longer to become as large and vibrant as it is today.

After Heat, I turned my attention to solving what I perceived as a technical gap in OpenStack: at that time, OpenStack had no functional interoperability with Kubernetes. This led to my involvement in contributing technically to the OpenStack Magnum project by writing a majority of the initial implementation and recruiting a talented core team from across the OpenStack community. Magnum provides technical interoperability by running Kubernetes on top of OpenStack in an OpenStack-native way.

Finally, I started the OpenStack Kolla project to solve OpenStack’s last significant pain point. Operational expenses quantified as the warm blooded people responsible for operating an OpenStack cloud were greater than other competition in the IaaS platform marketplace. OpenStack prior to Kolla required very large teams to maintain. With a Kolla-deployed OpenStack, this problem no longer exists. I passed the baton as the Project Team Lead in November 2016 to a solid leadership team I had developed in the 3 years Kolla was in development.

One thing we tried early on with Kolla was deploying OpenStack on top of Kubernetes 0.9.7;  in fact, that was the original goal of Kolla. At the time, Kubernetes was in its early stages of development and could not serve the complex use case scenarios OpenStack presented. Instead, we went through numerous iterations and eventually settled on Ansible as the basis for Kolla. Kolla works so well because Asnbile is such a great dependency choice and the Kolla community stuck with the implementation until we reached critical mass and success.

In March 2016, the core reviewer team of Kolla had made a determination that Kubernetes was mature enough to revisit. As PTL at the time, I didn’t have attention to dedicate full time to growing the community around an OpenStack-on-Kubernetes deployment tool. The core reviewer team started the work with my encouragement.  Today, the kolla-kubernetes core reviewer team has implemented a fully functional helm-based microservices layer. We have proven that OpenStack will run on top of Kubernetes, and there now are real, production deployments of the kolla-kubernetes deliverable of the OpenStack Kolla project.

The main factor that makes OpenStack great is the large, diverse set of community members working on the overall system using the Open Source methodology. The properties that makes VMWare and AWS great from the point of view of those respective vendors is vendor lock-in and lower operational expenses. The teams I’ve led in OpenStack have solved various aspects of the lock-in problem, and now with Kolla OpenStack has much lower operational expenses than the competition (AWS and VMWare).

There is no going back in life; instead there is only forward. To this end, the forward-looking strategy I believe is most crucial for OpenStack is competing against these closed source behemoths in cooperation with other Open Source Foundations.

As a result of my belief in this strategy, I am running for the OpenStack board of directors as an individual board member to drive Foundation Interoperability.  The first foundation I have a strong desire to build bridges with is the CNCF.  CNCF is hosting the Kubernetes, Helm, and other open source projects, of which I personally believe in harmony with OpenStack will lead to the best possible outcome against our closed source competition.

I encourage you to spend the few minutes to vote for the individual candidate that you believe best serves your desires for a better future for the world of software.  You may view all of the individual director candidate profiles or inspect mine in particular.

I believe without question that collaboration and foundation interoperability is key to both the success of OpenStack and CNCF.  Without excellent working relationships between these two foundations, the worst possible outcome may occur: competition between Open Source Foundations.  The world of Open Source should stand together rather than apart.  When you vote for the board of director membership, vote for candidates that will deliver on the best outcomes for OpenStack.


Steven Dake