15 April 2024

KubeCon / CloudNativeCon Europe 2024

Once again a KubeCon has come to an end

This year, the CNCF hosted the meeting in the French capital. In the south of Paris at the PARIS EXPO PORTE DE VERSAILLES, over 12,000 participants came together again this year.

In addition to various talks and unconferences, there was also a wide range of Contribfests slots, which made it easier for interested parties to get involved in various projects. The thematic focus this year was clearly on Artificial Intelligence and Machine Learning.

Wednesday morning started at 9 am. All keynotes had one thing in common: the topic of AI/ML! Here are a few titles: “Accelerating AI Workloads with GPUs in Kubernetes”, “Build an Open Source Platform for AI/ML” or “Optimizing Performance and Sustainability for AI”. In addition to the keynotes, participants were also offered plenty of talks on the topic of AI and ML. Anyone interested in this topic could find what they were looking for at any time during the three days in the “Paris” room. The new white paper on “Cloud Native AI” was also announced during the keynotes.

Location

Anyone wishing to enter the conference first had to pass through security, consisting of metal detectors and possible bag checks.

Unfortunately, there was again the problem of overcrowded rooms this year. Several times during the day it was necessary to reschedule spontaneously, as the selected talk was unfortunately already overcrowded and admission was no longer possible. Unfortunately, this also applied to some keynotes in the morning.

A few impressions

The CNCF Storage TAG and the Storage SIG reported on current developments in the storage sector and mentioned various white papers (“CNCF Storage Whitepaper, Performance and Benchmarking whitepaper, Cloud Native Disaster Recovery whitepaper, and the Data on Kubernetes whitepaper.”). Among other things, it was shown how “PersistendVolumeAttributes” can be used to adjust PersistentVolume attributes, for example to adjust the number of IOPS for a volume during operation.

In the Contribfest slot on Metal3 (Metal Kubed), the maintainers of the project gave a first insight and showed how a development environment can be set up. Among other things, Metal3 offers a ClusterAPI (CAPI) implementation that can be used to manage bare-metal systems. Ironic, which originates from the OpenStack project, is used in the background.

At the “From UI to Storage” talk Thanos maintainers gave an insight into the current implementation and potential future improvements.

In the “CRI-O Odyssey”, the CRI-O maintainers talked about innovations within the Container Runtime. This included the topic of “Confidential Containers” and “Podman-in-Kubernetes”. The topic of WASM integration was also on the agenda.

The maintainers of Fink reported on “Fink on Kubernetes” and how the system is used to classify objects like astroids or supernovas in the field of astronomy.

In the talk “eBPF: Abilities and Limitations”, not only general misconceptions were clarified, but also ways to work around existing limitations were pointed out. In addition, the question of whether eBPF touring is complete was raised and a version of Conway’s “Game of Life” in eBPF was presented.

Those who have always wanted to know how Istio mTLS can be implemented in multi-cluster environments with the help of SPIRE found what they were looking for in room D. After a short introduction to SPIFFE and SPIRE, it was shown how Istio components can be connected to the SPIRE agent.

Maintainers of operator-sdk talked about current innovations in operator-SDK and OLM V1. Particularly interesting were the changes in the Operator Lifecycle Manger that have been introduced in version 1 (previously version 0). Among other things, OLM v1 is now able to manage packaged operators via Helm, even without prior preparation by the operator maintainers.

Conclusion

In addition to many interesting talks, there were plenty of opportunities for lively discussions with other conference attendees and exchanges with exhibitors or project maintainers.

We are already looking forward to KubeCon 2025 in London!

Categories: Events
Tags: CloudNative CloudNativeCon CNCF Container KubeCon Kubernetes Paris

AV

About the author

Adrian Vondendriesch

Technischer Leiter

zur Person

Adrian ist seit 2013 Mitarbeiter der credativ GmbH. Als technischer Leiter des Cloud Infrastructure Teams beschäftigt er sich hauptsächlich mit der Planung, Realisierung und Betreuung verteilter Infrastrukturen wie zum Beispiel Kubernetes und Ceph sowie mit der Erarbeitung von Deployment-Strategien. Zuvor war er Teil des Datenbank-Teams bei credativ und war dort unter anderem mit dem Aufbau und der Verwaltung von hochverfügbaren Datenbank-Systemen betreut. Seit 2015 beteiligt er sich aktiv am Debian-Projekt.

View posts


Beitrag teilen: