Blog
Building CMLv2 Labs with Python
A programmable lab turns network learning into a repeatable experiment instead of a fragile collection of screenshots.
Rahul Nangare · August 24, 2026 · 1 min read
A programmable lab turns network learning into a repeatable experiment instead of a fragile collection of screenshots.
Describe the topology
Define nodes, interfaces, images, and links as data. A script can then create the same lab repeatedly and expose accidental differences.
Separate build from test
Provisioning proves that objects were accepted. Independent checks should confirm boot state, addressing, adjacency, and the behavior the lab exists to study.
Keep experiments cheap
Small topologies start faster, fail more clearly, and are easier to share. Add complexity only when a specific question requires it.
Field rule
A good lab is disposable infrastructure with durable documentation.