What is CTAN?

The acronym stands for “Comprehensive TeX Archive Network”, which more-or-less specifies what it’s for:

The basic framework was developed by a TUG working group set up to resolve the (then existing) requirement for users to know on which archive site a particular package might be found.

Actual implementation offers three distinct types of host:

Core archives
Which form a small, tightly-coupled set of machines, which perform management functions as well as serving files;
Mirror archives
Which do no more than take regular copies of core archives, and serve them; and
Archive selector
Which is a meta-service, which routes requests to an apparently “local” mirror (“local” is determined by an algorithm that uses your net address to determine where you are, and then selects a mirror that’s close).
Note that there is nothing to prevent any archive from supporting other functions, so a CTAN mirror may also operate as a CPAN (Perl) mirror and as a SourceForge (general free software) mirror, and …

Functions that distinguish core archives are:

Users may make direct contact with the CTAN management team.

Users should ordinarily download material from CTAN via the archive selector: this uses the mirror monitor’s database, and uses the caller’s geographical location to offer an efficient choice of “sufficiently up-to-date” mirror site for you to connect to. This procedure has the advantage of distributing the load on CTAN mirrors.

Note that all the download links, given in the web representation of these FAQs, are set up to use the mirror selector.

This answer last edited: 2013-05-21