For many years I thought of a subnet as a single contiguous range of ip addresses expressed using CIDR notation.
However on GCP a subnet can have a primary ip address range from which VMs get assigned an ip address a bunch of and secondary ip ranges that are used as ip aliases. For example GKE use one range for the kuberenetes pod network and another one for the kuberenetes service network.
- In TCP/IP can a subnet have more than one CIDR
- Is the definition of subnet with mutiple CIDR ranges in GCP unique to GCP or a common across others products and services?