Innovation and the pace of that innovation were the two main discussion points at last week’s KubeCon EU. The official community conference of Kubernetes. KubeCon was a platform for some important announcements including Kubernetes involvement in CNCF (Cloud Native Computing Foundation) for its governance going forward. This transfers control of Kubernetes to a wider committee, while still ensuring it stays razor focused on its direction and mission.
Attending the conference showed that Kubernetes is moving at an incredibly fast pace. Hidden features were revealed alongside announcing the project’s roadmap. There is so much innovation in the project from vendors and individuals demonstrating the bet many are placing on Kubernetes to become mainstream in the future.
Not only is the Kubernetes project innovating at pace, but also the developer eco-system around it is adding value on top. I’ve been impressed with a number of products that are doing amazing things to make Kuebrnetes an even more impressive platform. However, the challenge with a community project, and one that I saw last week, is that it is always changing. The opportunities developers see to add value, quickly become the project’s core features. On my blog, The Infrastructure Standpoint, I wrote that:
“…in the end, this is a community driven project and it’s the community that decides what should fit within Kubernetes and what should be left to someone else. That’s why it’s so important to be involved in the community on a day-to-day basis, to know what’s being built and discussed. When I asked Shannon Williams, co-founder of Rancher Labs, how he copes with this problem, he said you have to move faster, when part of your code is no longer required, just deprecate it and move on. Sure thing, you need to know how to move *that* fast, though!”
The Real Need for Kubernetes
Innovation and the speed of that innovation are essential in today’s dog eat dog world. For me though, hearing about the innovation is one part, seeing the real need for it is another. At KubeCon, we heard from the New York Times using Kubernetes in production. Eric Lewis showed how the New York Time’s were previously giving developers a server, to enabling developers to provision applications using Chef, to containers with Fleet and now Kubernetes. According to them, that’s definitely the best thing to deliver developers’ infrastructure at present.
However, Kubernetes is capable of handling only parts of an application. In fact one end customer told me that a great orchestration software should be able to handle both containerized and non-containerized workload. That is exactly what we are doing within Flexiant Concerto. It is capable of automating the deployment on traditional virtual machine, by means of configuration management as well as handling container-based workload powered by – of course – Kubernetes.
Pace of Change Equals Caution
The problem with this pace of change is that it creates questions on which approach to use – what horse to bet on, what if the selected standard is the one that gets deprecated? While competition is good, it drives a totally understandable caution from end customers. I kind of miss the time when the standard was coming first and products were based upon them, but now we tend to welcome de facto standards instead, which take some time to prove their superiority.
What really matters is having more people using Kubernetes. More use cases will drive more innovation and will bring that stabilization required to convince even the most cautious ones.
You can see how Flexiant is adding value to Kubernetes by watching our Flexiant Kubernetes Orchestration as a Service demonstration. Kubernetes Orchestration as a Service allows you to get Kubernetes on any cloud, easily leaving more time to focus on the creative side of Kubernetes. This video shows you how easy it is to launch the ‘hello world’ guestbook feature in Kubernetes.