The future of infrastructure as code

Before we can foresee the future, we need to understand the past.

The term infrastructure as code is not new. Its roots are deep in the '90s, when sysadmins used Perl and Shell scripts to automate their regular routines. Sysadmins were the same as developers, but with a different focus (mostly automation and the stability of their infrastructure).

This focus started to shift as soon as Amazon Web Services was launched in 2006, and it became a matter of a few minutes to have a fleet of virtual machines at your disposal. The increased speed of provisioning introduced the issue of scaling performance.

In addition, it was hard to develop your own infrastructure management tools using high-level programming languages. In the end, configuration-management tools started to appear: Chef, Puppet, Ansible, and so on.

Later on, idempotence and declarable infrastructure became necessary requirements. Configuration-management tools could provide...