In cloud native environments it's common to load balance requests between a number of containers that make up a service. eBPF can be used to build a highly efficient load balancer that runs in the kernel.

This talk starts with a quick introduction to make sure everyone understands what eBPF is, and a brief primer on how IP and TCP packets are constructed so that the audience is clear on what a layer 4 load balancer needs to do. Then learn how an eBPF Load Balancer works to redirect packets to and from different containers as you see a basic load balancer written from scratch in a few lines of code.

You don't need to know anything about eBPF to follow this talk, nor do you need to be a networking afficionado. However, it does assume that the audience has a basic familiarity with (Docker) containers and IP addresses, and enjoys watching live-coding!