Cloud Native, Ahoy!

After attending ContainerDays for the third time within a year, I think it's fair to say I'm a pretty big fan of the series. This year's Hamburg edition was once again a blast, and I wanted to share my impressions and some insights from a more technical perspective. My focus will be on the trends, topics and talks that stood out to me across AI infrastructure, Platform Engineering, Kubernetes and cloud native software engineering.

The official program featured 120+ sessions across six stages, so there was plenty to choose from. I wrote a recap half a year ago after attending ContainerDays London.

A small disclaimer: this recap will, of course, be a biased one. There's simply no way to attend every talk across every track, and there were some areas I found more interesting than others.

I also won't talk about the venue, location or organization itself. You'll have to find that out by attending yourself (I definitely recommend) 😉.

And one last thing: the talks were recorded and are set to be published on the ContainerDays YouTube channel, so you can catch up on the sessions you missed (incuding mine 😊).

Have you ever heard of AI?

The actual full name of the conference was ContainerDays & AI Context, so Artificial Intelligence and AI-Infrastructure were naturally among of the main topics of the event. A whole track was dedicated to AI, and plenty of exhibitors were showing AI-based solutions.

The most common use cases I came across were AI-assisted incident handling, including root cause analysis, and AI-based cluster optimization, such as dynamically rightsizing workloads to reduce infrastructure costs.

Personally, I attended several talks around infrastructure for model serving and AI Workloads on Kubernetes. The two I particularly enjoyed (and also managed to join 😂) were:

"From Minutes to Seconds: How we reduced ML models cold start time by 90% on Kubernetes"

by Simone Sciarrati & Federico Hernandez, and

"But we have GPUs at Home - On-Prem AI with Kubernetes, OpenStack & KubeVirt."

by Maik Zimmermann & Nicolai Ort.

However, the most memorable session for me personally was:

"From Von Neumann to Transformers: An Infra Engineer's Guide to What's Actually Inside an LLM." (Stefan Schimanski & Sergiusz Urbaniak)

Even though my brain got absolutely fried during the session (30 minutes was way too little time to explain such a nuanced topic), I still found it extremely interesting 😄

Platform Engineering to the Rescue

Luckily, the conference featured six stages in total, so there was plenty to choose from beyond AI.

Having built multiple developer platforms throughout my career, I was especially interested in the Platform Engineering track.

Two recurring topics were reducing ‘Platform Obesity’ and improving existing Internal Developer Platform setups. I particularly enjoyed Joep Piscaer’s take on "Platform Obesity, not Complexity, (…) killing our platforms."

The main point that stuck with me was pretty simple: good platforms need continuous feedback from the people actually using them, and platform teams shouldn't be afraid to say "No" to feature requests when they add unnecessary overhead.

Running platforms at scale was another area I checked out. I joined Simon Bein’s

"Scaling Kube-Apiserver to Thousands of Tenants - Lessons Learned From kcp."

and

"Not All Managed Kubernetes Are Created Equal: A Deep Dive into Control Plane Architectures."

by Artem Lajko & Nick Berthold.

I really appreciated seeing these real-world journeys in action. Learning fresh concepts and tools is always a win, especially with how fast this domain moves.

Good ol' software engineering

Finally, as I have to build software from time to time (and am very happy about it), another track I kept coming back to was Cloud Native Experience & Golang.

There were several sessions around Kubernetes Operator Development, which is especially relevant when building and extending Internal Developer Platforms. I checked out

"Building Production-Ready Kubernetes Operators."

by Fortune Ndlovu,
as well as

"Beyond Reconcile Loop - Patterns for Kubernetes Operator Development

by David Steiman.

Natalie Godec delivered a good ol’ migration talk that I really enjoyed:

"But can you really run your app on 2 Clouds at the same time?"

I definitely picked up a few new ideas from those sessions that I can take into future projects.

A little WebAssembly on the edge

I was also lucky enough to be back on stage myself this year. Thanks to everyone who joined my session

"Kubernetes on the edge: running low-latency workloads with WebAssembly."

Special thanks to all of you who asked the questions and reached out afterwards!

In this session, I looked at WebAssembly as an alternative runtime for selected Kubernetes workloads, including the architecture of a WASM-compatible Kubernetes cluster, the configuration needed to run WASM-workloads and, of course, a live demo.

The interesting part for me is not replacing containers everywhere. It is understanding where a smaller runtime footprint, fast startup times and low-latency requirements make Wasm worth looking at alongside traditional containers.

Once the recording is available on the ContainerDays YouTube channel, I'll add it here and share on shiftavenue’s LinkedIn.

See you next year?

As written in the intro, the overall experience was exceptional and the organizers did a great job once again. I can definitely recommend checking out a ContainerDays event if cloud native, Kubernetes or Platform Engineering is somewhere in your world.

The series has grown well beyond Hamburg by now, which is probably a good sign that I'm not the only one enjoying it. 😉 ContainerDays has already expanded to London and Singapore as well.

For my part, I will definitely keep ContainerDays on my conference agenda next year. See you there?

Cheers and have a good one,
Rene