Conform gives way to cfengine

For over twenty years the configuration of CSE's servers and lab computers has been managed by a home-grown product called conform. conform micromanaged the entire filesystem of each computer and handled the copying in of all configuration files, binaries, scripts, libraries, data files, etc, from a central repository called "the conformary" (and NFS server). conform is very powerful, but is also very labour intensive in regards to maintaining the conformary and the thousands of specification files which determine conform's actions. Also, being so old conform lacks many newer ideas such as pre- and post-install scripts. conform is part of the Old World.

The New World grew out a HoS-supported project to move as much CSE infrastructure as possible out of CSE's K17 data centre and into "the cloud", specifically Amazon's AWS. Early on, a decision was made that conform would be left behind and that all new servers and hosts set up in AWS would be managed some other way. This other way ended up being -- after a review of configuration management tools -- a split between cfengine and Debian's apt family of package manager utilities (apt, dpkg, apt-get, apt-file, etc.) In this split, apt handles all Debian package management (mainly installs and updates) and cfengine handles the rest, mostly configuration files.

Debian was chosen because CSE's existing teaching platform was heavily based on Debian and the principle of least surprise was applied.