In this interview from KubeCon Paris, 2024, Stephen Hoekstra describes how he and Schuberg Philis were inspired by Google to implement Cilium as a data plane. Hi, my name is Stephen. I am a mission-critical engineer working at Schuberg Philis. We are based in the Netherlands, just outside Amsterdam, and we look after mission-critical or vital industries in the Netherlands. We focus on the mission-critical part, so we started out with the AWS VPC CNI. We weren't quite satisfied with that, had some teething issues with Calico, and then randomly we came across a blog post from Google talking about how they had just picked up Cilium to use in their data plane in GCP, and I joked to a colleague, "If it's good enough for Google, it must be good enough for us." With that in mind, on the Monday we went to the website, picked up the getting started guide, and within half an hour we had a proof of concept up and running. From there we reached out to Isovalent, told them about our interest, described what we do, and had some more concrete ideas of value to see what we could do with it, and what more value we could get out of it than just the CNI. They basically helped us from the start, talking about default-deny policy. Out of the box, your containers or your application can do nothing, and then you can add rules to say this can talk to that, but can't talk to this. The problem for us was that we also didn't want all the legwork of configuring this for developers, so when they showed us the Hubble UI, this was a great way to say to our developers, "Hey, you can deploy something. It may not work out of the box the first time. Look at Hubble UI. You see all the traffic flows, you see the network map, and it's very easy to identify what the problem is. Click on the link for a suggested policy to apply." That really helped the onboarding of integrating Cilium into our platform and making the developers pick it up, and it made it successful.