5 Takeaways from KubeCon Europe 2019

 

by

KubeCon and CloudNativeCon is the world’s largest conference about Kubernetes, the surrounding ecosystem, and the principles of Cloud Native. Container Solutions sent a team of engineers to Barcelona in late May in order to connect, learn, and explore with the larger Cloud Native community. It also presented an amazing opportunity to better connect with our team itself, as we are based in offices all over Europe and North America.

Photograph of Container Solutions engineers Philipp Strube, Adam Otto, Piotr Perzyna, Thomas Richter, Andreas Krivas, Luke Heidecke, Riccardo Cefala, Bogdan Lordache, and Rafael Portela connect at KubeCon Europe 2019


Container Solutions engineers Philipp Strube (from left), Adam Otto, Piotr Perzyna, Thomas Richter, Andreas Krivas, Luke Heidecke, Riccardo Cefala, Bogdan Lordache, and Rafael Portela connect at KubeCon Europe 2019.

There was a lot to take in at KubeCon Europe this year. We heard many interesting talks (sometimes at conflicting times), met vendors, and learned about new developments in basically all areas of the Cloud Native world.

In order to keep it concise I wrote down my top five takeaways of KubeCon Europe 2019:

Operators are the new thing.

As the demand for stateful applications running in Kubernetes become higher and higher, Developers and Operations alike are facing challenges in managing and deploying these applications in Kubernetes. At KubeCon there were a few interesting talks addressing these topics with operators. This seems to be the new thing after service meshes, and we may hear much more from operators in the near future as vendors and the open source community expand their support.

An operator has specific operational knowledge about the applications that run in Kubernetes, and it is a way to interact and manage the application. The most straightforward operators are the ones that can deploy and manage stateful applications. For example, they may provide answers on how to backup a database periodically and automatically, reducing operational overhead while automating cluster operations—one of the main fundamentals of Kubernetes. Because an operator is implemented only by using Kubernetes-constructs, it is independent from the cloud provider.

New interface alert: Cluster API.

Like we manage pods with kubectl inside a cluster, we may want to install, manage, and configure multiple clusters in a similar fashion. As we well know, this can be a tedious task. This is where ClusterAPI comes in, a non-cloud provider specific interface, implemented with a management cluster. It Is part of the Kubernetes SIG Cluster Lifecycle community project.

Instead of managing pods, the Cluster API can interact with Machines. Therefore, similar to ReplicaSets for Pods, we have MachineSets. And so, of course, we also have MachineDeployment, an analog to regular Kubernetes Deployments.

Cloud-provider specific features can also be used, referenced, and easily applied to the cluster.

Service meshes are getting easier to implement.

Services meshes are so hot right now! In a microservices application, managing networking endpoints can get complicated and difficult to handle very fast. Using the service mesh pattern helps to make networking much smarter in a Kubernetes cluster.

In order to standardize the implementations of service meshes, the Service Mesh Interface (SMI) has been announced. Therefore, managing and installing a service mesh of your choice should be much easier in the future. The implementation of the service mesh itself will be realised by service meshes such as Istio and Linkerd.

Helm is going through some changes.

No more Tiller with Helm 3! And various other changes which may or may not make Helm 3, the latest version of the Kubernetes package manager, feasible to use, like replacing Go as a scripting language with Lua.

I don’t want to recommend any specific Kubernetes package manager. It was, however, certainly interesting to see the interest in Helm 3 at KubeCon.

Kustomize is natively included in kubectl. However I couldn’t find any talk about Kustomize at KubeCon. A lot of attention is now on Helm 3, for better or worse, and everyone has to decide individually which solution may fit their purpose.

The open-source community continues to be awesome.

For me, the most impressive thing at KubeCon Europe 2019 was not a piece of technology. It was the amazing open-source community. With more than 7,500 attendees, there were hundreds of talks and sessions, get togethers, meetings, and parties.

The community is in a very vibrant and healthy state and a lot of people want to contribute in some way. And with the help of the Cloud Native Computing Foundation (CNCF), it provides many possible entry points to contribute: translating, improving documentation, approving pull requests, raising issues on GitHub, contributing to or just listening in on the many online sessions. The community frequently discusses different topics within its special interest groups (SIGs); for instance, the Network SIG discusses the future of the Kubernetes networking world. . A full list of all SIGs can be found here. It’s a great way to get started or dive deeper.

It is very easy to lose focus with all the things going on in the Cloud Native world. However conferences like KubeCon can not only be very helpful to expand your horizons, but can also help to figure out what is most important to you.

For the people who didn’t get enough at KubeCon or the ones who couldn’t attend: The CNCF is providing a lot of talks on their YouTube Channel. Simply search for the topics you find most appealing and take notes!

Looking for a new challenge? We’re hiring!

C

Leave a Reply

Your email address will not be published. Required fields are marked *