Alex Jones avatar
By Alex Jones


Learn how to communicate between pods on different nodes in Kubernetes, including the journey of data through the cluster network and implementation-specific details.



This video is about node-to-node networking in Kubernetes, specifically how one pod can communicate with another pod on another. Before we start, let's recap what we learned last time about how we connect between pods on the same host. Firstly, the pod sends data across the veth pair and over to the bridge. If the bridge app fails and there isn't a connected device with a corresponding MAC address for the IP data that's come in through the data frame, then it will go along the default route. The default route is eth0 in this scenario.

How does node-to-node networking take place?

The IP address trying to be hit from the origin pod and the destination will be resolvable and understood in the cluster network topology. This is a bit nebulous in this video because this is implementation-specific. We will talk a little about the implementations in a subsequent video around CNI or cluster network interfacing.

Once the data has gone through the cluster network and has been correctly sent to the corresponding node, it will be the same journey on the inverse where it will be received through its ethernet device. It will then go through the same journey for the veth pair across the network, a namespace, and then into the target pod and container, which is how this journey is completed.

In the next video, we will discuss cluster network namespacing, cluster network interfaces, and how this can be implemented across cloud and on-premise providers. Thank you.

Don't stop now, check out your next lesson