01
6.1 Impact of automation
- Benefits: consistency, speed, scale, fewer human typos. Risks: a skills gap and a larger blast radius when an error propagates.
- Automation ≠ orchestration (orchestration coordinates many automated tasks). Config drift = devices diverging from intended state.
02
6.2/6.3 Controller-based & SDN
- Traditional = each device has its own control plane. Controller-based = centralized MANAGEMENT (e.g. Cisco Catalyst Center, formerly DNA Center). The data plane stays distributed.
- SDN planes: control plane decides (routing), data plane forwards. SDN centralizes the control plane.
- Northbound API = up to apps/users (REST). Southbound API = down to devices (NETCONF/RESTCONF/OpenFlow). Overlay (virtual/VXLAN) vs underlay (physical) vs fabric (both managed together).
03
6.4 AI/ML in network operations (NEW in v1.1)
- Predictive AI: anomaly detection and forecasting (e.g. predicting failures/capacity).
- Generative AI: produces configs, summaries, assistant answers. ML is the underlying technique.
- Explain-only - no code or math required, but it is now an exam topic.
04
6.5 REST APIs
- Stateless HTTP/HTTPS. CRUD → verbs: POST=Create, GET=Read, PUT/PATCH=Update, DELETE=Delete.
- Status families: 2xx success, 4xx client error (401 unauthorized, 404 not found), 5xx server error.
- Data encoding: JSON or XML. Auth types: basic, token/API-key, OAuth (new in v1.1).
05
6.6 Config management (Ansible/Terraform)
- Ansible: agentless, push model, over SSH, YAML playbooks.
- Terraform: infrastructure-as-code, declarative, HCL (new in v1.1, replaced Puppet/Chef in the blueprint).
- Concepts: agent vs agentless, push vs pull, idempotency (re-running yields the same state).
06
6.7 JSON
- Objects {} hold key:value pairs; arrays [] hold ordered values. Strings use double quotes.
- Types: string, number, boolean, null, plus nesting. No trailing commas, no comments.
- You read/interpret JSON on the exam - you do not write it.