2016 was my first OpenStack Summit and my eyes were opened to something new and fresh that I had not experienced before. With my closest 8,000 friends I got to hear from real OpenStack users, talking about the problems they were trying to solve and their experiences along the way. This year was much of the same content, but with around 3,000 fewer attendees. I don’t attribute the decreased attendance to a lack of interest; rather, this summit came with some challenges in travel and expense as well as competition from other conferences. Regardless, the ecosystem and community is alive and pushing forward to the common goal of creating an open infrastructure to run cloud services on — and it is exciting.
Many of the keynote speeches focused on the maturation cycle of the community. As OpenStack has now hit the 7-year mark, it is time to start growing up. One of the first topics covered was the perception versus reality when talking about OpenStack success. The perception of the Twitter-sphere is that the project is dying and no one is using it. The reality, however, is that OpenStack manages 5 million cores in production and boasts an ever-growing list of companies, including GE, AT&T, Ebay, and OVH, that are using it to provide services either internally or externally.
Another topic that was discussed was the future of OpenStack. There seems to be a perception that OpenStack was going after and competing against AWS; however, my takeaway from the keynotes and sessions I attended was that OpenStack’s primary use case is in managing and providing a single control plane as we start to build private clouds everywhere in an edge or fog cloud computing use case. Personally and professionally I am curious to see how this develops, as I think this is going to be part of the next big phase of computing when we begin to decentralize the cloud as we know it today. Latency is something we can’t get around and data gravity is forever increasing. Maybe Azure actually knows what they are doing with releasing Azure Stack?
OpenStack’s success is also dependent upon openness to incorporate other open projects instead of reinventing the wheel every time. There is a finite amount of people working on open source. It makes sense to create standards across all open projects to increase adoption and provide long-term feasibility for the companies that are early adopters.
And, it was hard not to notice the focus on Kubernetes. Kubernetes was mentioned in almost every session I attended and was a common topic in each of the keynotes, as well. There is potential to use Kubernetes to run the OpenStack infrastructure, and then use Kubernetes to run your cloud applications. Eric Wright’s presentation does a great job of breaking it all down: Kubernetes on OpenStack on Kubernetes: The Infrastructure Club Sandwich. This is an interesting development as containerizing the OpenStack projects will start to solve the day 2 issues that occur: upgrades, versioning, and dev environments will all potentially become easier with Kubernetes.
OpenStack Summit has posted all of the keynotes and sessions online, if you are interested in learning more about open source infrastructure: https://www.openstack.org/videos/. And be sure to check out my 10 minute lightning talk on Expedient’s OpenStack project: