(to see pictures from the event (and to add your own), go here)
A little late in writing this, but I thought that I would send out my write up of KubeCon EU. If you have written a write up or have found one that was insightful, please post the link. Also, if you have a favorite talk, please add the link, or at least the title of the talk, so that we all can view it. (All opinions are my own.)
The first event of the week was Rejekts, which is one of my favorite events. KubeCon gets thousands of submissions and the best of the ones not selected, for whatever reason, end up at Rejekts. And since it is so much smaller than KubeCon, you get exposed to a wider selection of topics, like “How eBPF Actually Works”, “Burnout++ - Recognizing and Managing Burnout”, and others about Fluent Bit, Cilium, Uber talking about how they migrated 2 million CPU cores, etc. I will say, the one that caught my eye first was “Will ARM be the new Mainstream in our Data Centers?”. It was a great talk and I spent an hour afterwards speaking with the speaker Tobias Schneck.
If you are able to go to KubeCon NA, come a little early and attend Rejekts. It was a lot of fun.
(full disclosure- my talks were rejected by Rejekts, so I think we should create t-shirts that say “Rejekts- Rejekt” or Rejekts ^2)
KubeCon EU
The day zero events are the mini conferences. Unfortunately, as more of generalist in the Cloud Native world, there was too much to choose from. I did get to see the WASM event. They led off with the question, “Is this WASM’s Docker moment?”, meaning is WASM going to take off this year. I also was able to see Kubernetes on Edge Day. This is a big thing in the arm64 world because as edge devices gain in power, they can do more compute on the edge, without sending data to the cloud. Therefore, they can do more.
On the first day of the event, you realize how big it is with over 13k attendees. It is one of the largest events in Europe. The theme of day was AI. NVIDIA started of the day talking about how everything should be done on GPU. This is not a good thing for the environment, given how much more energy GPUs use than x86 chips do. Fortunately, the next talk was a panel discussion titled, “Optimizing Performance and Sustainability for AI”. Community member @ppouliot (Peter Pouliot) was on the panel talking about how GPUs are not needed for AI Inference.
The second day’s theme was sustainability, which was really nice to hear. Most of the keynotes talking about how they are improving energy efficiency. For example, Scaleway talked how they are moving to more arm64 servers to save energy and how they have a datacenter that has a Power Usage Efficiency (total power/power used for compute) of 1.18, where the average is 1.58. Next was the introduction of SPIN a new project. The teams mentioned that they were able to reduce the power consumption of by 60%. Even Deutsche Bahn talked about they are better provisioning servers so that they are not running when they don’t need to be.
For the show floor, Ampere didn’t have booth, but my colleagues and myself were all over the floor speaking. I spoke at the Scaleway booth, and we had speakers at the Redhat, SUSE, Canonical, Equinix, and Oracle booths.
Overall, the event was great, interesting, and huge like the other KubeCons that I have attended. What made this one a little different was that it was in Paris, which does make it a little tougher to get approval from management. They don’t hear that you are going to KubeCon, they hear that you going to Paris. 
And as a closing note, here is a fun little conversation about the best swag at the event.