<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Cilium :: Tag :: Rejekts2026 and KubeCon EU 2026</title>
    <link>https://kubecon25.nicolai-ort.com/tags/cilium/index.html</link>
    <description></description>
    <generator>Hugo</generator>
    <language>en</language>
    <atom:link href="https://kubecon25.nicolai-ort.com/tags/cilium/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Unleashing the tides of kubernetes networking by removing kube-proxy</title>
      <link>https://kubecon25.nicolai-ort.com/day-2/06-kubeproxy/index.html</link>
      <pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate>
      <guid>https://kubecon25.nicolai-ort.com/day-2/06-kubeproxy/index.html</guid>
      <description>Code/Demo A talk by isovalent (now part of cisco - god i love that they have to say this every time). It’S a good baseline introduction to how kubernetes service routing works but also a bit dry (in terms of the presentation itself). I skipped the introduction to cilium in these notes. The docs exist for a reason.&#xA;Kubernetes Services - a baseline East-West: ClusterIP -&gt; App2App inside the cluster North-South: NodePort -&gt; External Client to app in Cluster Kube-Proxy - IPTables Mode IPTables: Traffics flows through different tables/chains - most imporantly the NAT-Table Every Node has it’s own kube-proxy next to the kubelet ClusterIP: Scales to a huge numer of rules when exposing multiple services NodePort: Masquerades sources if routing cross-node (Source-IP is lost) TODO: Steal iptables visualizer TODO: Steal livecycle of a packet clusterip TODO: Steal livecycle of a packet nodeport</description>
    </item>
  </channel>
</rss>