Bosonit Experiences: CNCF Europe 2022
Hello, I am Jon RodriguezI am part of the team DevOps and Systems. I have been working for Bosonit for several years and I am fascinated by technology. But especially automating processes. In recent years there has been a tendency to put applications, which used to be hosted on the machines themselves, into containers. But do we really know the limits of containers?
To this end, I am here to talk about my experience at the conference. KubeCon + CloudNativeCon Europe 2022 (CNFC) and some of the technologies that seemed to me to be more innovative.
What is CloudNativeCon?
This is a conference organised by the Cloud Native Computing Foundation (CNCF), which is part of the Linux Foundation. The CNCF is in charge of incubating and maintaining Free Software projects designed to run on Cloud environments. It currently comprises more than 120 projects: Kubernetes, Prometheus, Conteinerd and Harbor, among others.
Why should we attend such events?
It is true that at events of this magnitude, a few days after the end of the event, the lectures are uploaded on YouTube. It seems absurd to spend a lot of money and time to go and watch these events. But the reality is that by being there we can interact with other spectators, ask questions, discuss solutions and meet very interesting people. It's very difficult to try all the open source solutions and to keep up with what's coming out and how they're evolving. But that's why you are there, they tell you about their use cases, new functionalities, tests they have done, for which use cases they use each solution.
What was your experience at KubeCon + CloudNativeCon Europe 2022?
The truth is that I had a great time, I managed to learn more about technologies that I had been wanting to learn more about for a long time and I met some amazing people. The day began with some presentations in the main room (keynotes), where they told success stories of companies adapting these technologies, technology improvements and new functionalities.
Then the solutions pavilions opened, where hundreds of companies had their stands and told you about the solutions they had developed. There was also a Free Software area where new project additions to the CNCF were on display. During the day there were presentations and talks about the technologies in detail, in different parts of the complex, where you could learn and go deeper into the technologies.
What are the limits of containers?
The limits of containers lie in the executing system. Running containers on Docker Swarm is not the same as running them on a Docker machine or on Kubernetes. Also, Kubernetes is a very customisable, allows modules to be replaced by others that follow the same standard.
These modules can be useful for scaling up the cluster event-driven, debug developments, have detailed monitoring, increase the capabilities of one of the components of the clusteramong other functionalities. One of the configurations they presented at the event, could run up to 1 million pods in a single clusterdistributed over 100,000 nodes.
Which technologies do you think are the most innovative?
In my view, the more innovative solutions that have surprised me are the following:
- Direktiv: This Open Source solution allows you to run event-driven containers both on-premises and in the Cloud. It is a solution serverlessas well as other solutions on the market such as AWS Lambda. Specifically, Direktiv's technical team indicated that it was far superior to Lambda and other competitors. It works like this: A call comes in to our API, a container is raised in a cluster Kubernetes, it executes the code we specify, and once it is finished, the container is destroyed. They gave me a demonstration of how it works and I was surprised by the result.
- Kubedge: It extends the capabilities of Kubernetes to edge computing. We may not be familiar with these concepts, but it allows you to run containers and perform data processing on low-resource devices. For example, in an environment with limited connectivity, data could be processed in-house, and even ML models could be applied. This would reduce network traffic, latency and processing overhead in the cluster main.
- Akri: It is also a solution designed for the edge. It extends the capabilities of Kubernetes by allowing auto discovery of video devices, GPU's and IP cameras. It detects nodes that have access to these devices and schedules tasks on them. They are adding other protocols and systems to integrate with. I think it's a very interesting solution for environments with devices. IoT.
I am looking forward to testing these solutions and hope to be able to do so soon. I encourage you to come to upcoming events and to share your experiences with the rest of your colleagues and the community.