concept Netfilter in category kubernetes

appears as: Netfilter
Core Kubernetes

This is an excerpt from Manning's book Core Kubernetes.

We’ve looked at a lot of primitives up until now. But there\’s one final nugget we saved till the end: BPF. The Linux operating system separates the notion of userspace from kernel space. Using technologies such as eBPF and LLVM, we can inject new behaviors into the kernel, which modifies existing data structures on the fly. The most concrete use case for this in current day Kubernetes is, of course, networking, where IPTables and IPvS (which are used by kube-proxy - to access Netfilter), technologies may be leapfrogged to be much faster, more scalable, and easier to debug over time.

As you’ve seen, iptables is a binary that allows a system administrator to configure rules in the Linux kernel. The part of the kernel that typically processes network traffic is called Netfilter, and iptable configures Netfilter. Netfilter has a total of five different hooks that can trigger various events during different stages of network traffic, and iptable tables are associated to the hooks.