Cumulus VX Spine and Leaf

After hearing the word Cumulus twice from different initiatives on the same day, I decided I wanted to know more about Cumulus Networks in general, and playing with VX seems to be a great start. I am already running Vagrant and VirtualBox for other means, so having an additional box is easy. Well, the idea was just an additional box but after doing some GitHub investigative work, found out that there is already a pre-defined Cumulus Linux Demo Framework or Reference Topology available for consumption. I quickly followed this repository and built my own spine and leaf architecture:

The whole process did not take more than 10 minutes. There is a lot that goes in the background, but still, not bad for a virtual non-prod environment or validation platform that supposedly has the same foundation as the Cumulus Linux and Cumulus RMP versions, including all the control plane elements.

The configuration is done on each of the VMs using the Network Command Line Utility (NCLU) or by editing the /etc/network/interfaces and /etc/frr/frr.conf files. This definitely requires some “essential” Linux skills. Multiple demos are available here using this topology, including NetQ. I have tested the config-routing demo and it worked perfectly with two spines, two leafs, and two servers. It uses an ansible playbook to push the configuration to the spine and leafs, as well as adding new interfaces to the servers for the connectivity test. A nice way to test the OSPF and BGP unnumbered concept. 

The fundamental piece is the FRR (Free Range Routing) responsible for EVPN, BGP, and OSPF functionality. Pete Lumbis did an excellent whiteboard session at Networking Field Day 17 by going over the building blocks followed by a demo on a similar topology running Cumulus VX.