BGP Prefix Aggregation in Isovalent Networking for Kubernetes 1.17 simplifies routing by summarizing multiple pod and service IPs into broader prefixes, reducing route table size and improving scalability. Today I'm going to talk about BGP route aggregation. If you're a network engineer, you may have used it in the past. If you're a DevOps engineer, it may be a newer concept, but it's not that complicated. When you connect your cluster to a network, we typically use BGP with Cilium. Cilium comes with a built-in BGP daemon that lets you announce IP addresses from your cluster into the rest of the network. Typically we announce Kubernetes service IPs individually, so one route per service, like a /32 for IPv4 or a /128 for IPv6. The more services you have, the more individual IP addresses you announce, and that takes space in the routing table of the underlying network. A cleaner way to do this is summarization or aggregation: announcing a broader network that includes all of those individual IP addresses. Let's look at the demo. We start with a cluster connected over BGP to a top-of-rack device, and it's advertising all of those individual service IP addresses to the rest of the network. On the top-of-rack device, I can see all of those routes being advertised, and it's starting to fill up the routing table. Aggregation can summarize those /32 routes into a larger prefix, for example a /24, which is what we're going to do here. I'm going to edit my BGP advertisement configuration and add the aggregation length, which is how I specify the subnet mask for the larger prefix I want to announce. Instead of individual IP addresses, we're going to announce a bigger range, a /24, which can cover up to 256 addresses. We also support IPv6, but I'm using IPv4 in this demo. When I connect back to my top-of-rack device, instead of a list of /32 routes, I now see a single /24. That makes the routing table much more compact and easier for the device to process. It's a very simple configuration: add one line to your BGP CRD and you're good to go. That's all for today. Thank you very much for watching, and I will see you in the next one.