Bootcamp: Cisco Catalyst SD-WAN Automation
Four hours of automation for engineers who already know SD-WAN. You will not learn what a TLOC, OMP or config group is — you will learn automation using those objects as the vocabulary. Together we build a Python toolkit against a real Manager: an authenticated session, an operational state snapshot, config groups with the asynchronous task model, Terraform, and a validated change pipeline with automatic rollback. You leave with the complete repository. Lab credentials come from HashiCorp Vault; no secret ever touches the code.
1Starting point: why automating SD-WAN is different
2 exercícios
Starting point: why automating SD-WAN is different
2 exercícios
Ten minutes to align the mental model and get the environment up. The difference between automating an SD-WAN fabric and automating a traditional network is not about tooling — it is architectural.
Controller-mediated automation
Environment up: repo, Vault and the offline suite
2Module 1 — Connecting to the Manager
3 exercícios
Module 1 — Connecting to the Manager
3 exercícios
45 min. The two-step handshake, written by hand with plain `requests`. After this the toolkit does it for you forever — but only once you have felt why it exists.
Authenticate: JSESSIONID and X-XSRF-TOKEN
The `data` envelope and the fabric inventory
Proof of life: how many edges are reachable?
3Module 2 — Operational state as data
3 exercícios
Module 2 — Operational state as data
3 exercícios
50 min. The central idea of the bootcamp: a change is only safe if you can prove the fabric came out the same or better. Proving it requires a comparable snapshot, as data.
Real-time or the statistics database?
Collect the state: BFD, control connections and OMP
Why the diff has an opinion
4Module 3 — Config Groups and the asynchronous model
3 exercícios
Module 3 — Config Groups and the asynchronous model
3 exercícios
50 min. Here you write to the fabric for the first time. Three disciplines arrive together: preview before deploy, wait for the task to finish, and namespace your work in a shared lab.
From templates to config groups
Preview: the dry-run almost nobody uses
Deploy and polling: the script that doesn't lie
5Module 4 — Terraform, properly
3 exercícios
Module 4 — Terraform, properly
3 exercícios
45 min. The same change as module 3, now declarative. The point is not that Terraform is better — it is that Terraform is only usable by someone who understands what it is hiding.
Read the provider schema (the error is deliberate)
Drift: the superpower the script didn't have
The uncomfortable discovery
6Module 5 — Validated change pipeline
2 exercícios
Module 5 — Validated change pipeline
2 exercícios
30 min. The capstone: everything you built, together. Snapshot before, apply, snapshot after, compare — and roll back if it regressed.