What does "infrastructure automation" mean, and what are the two ways a platform can be automated?
Using code to configure, provision and manage applications together with the compute, storage and network they run on, either through automation tools (Ansible, Puppet, Chef) or by calling the platform's API directly.
The key word is code. Instead of an admin clicking through consoles or typing commands device by device, the desired setup is written down in a form a machine can execute, so it can be repeated, reviewed and versioned.
The scope covers the whole stack, not just servers:
- the applications themselves,
- the compute, storage and network infrastructure under them,
- the network services (DNS, load balancing, firewalls) that connect them.
A platform can be reached in two ways:
- Through a configuration-management tool such as Ansible, Puppet or Chef. The tool knows how to talk to many platforms and gives you one language for all of them.
- Through a direct API on programmable infrastructure. The device or cloud exposes a REST (or similar) interface, and your own script calls it.
The tools usually sit on top of those same APIs, so the choice is between convenience and uniformity on one side and full control on the other.
Go deeper:
Infrastructure as code — Wikipedia — declarative vs imperative approaches and the main tools side by side.- Configuration management — Wikipedia — the discipline that Ansible, Puppet and Chef automate.