If you have the hostname, it's just a matter of string manipulation to extract the domain.

The trick is knowing what counts as a domain.

With gTLDs (.com, .net, .org, etc.), it's easy: you just want the last two atoms of the hostname ...

my $domain = join( '.' => ( split /\./ => $host )[-1,-2] );

... but with ccTLDs (.uk, .jp, .us), things get a lot more tricky as many of those have their own .com equivalents (.co.uk) which should be considered to be the real domains, or -- worse still -- use a mix of the two and three level styles.

An imperfect approach could be...

my @atom = split /\./, $host;

if ( length($atom[-1]) == 3 ) {      # gTLD

    $domain = join( '.', @atom[-2,-1] );

} elsif ( length($atom[-1]) == 2 ) { # ccTLD

    # Check if ccTLD domains are name.x.tld or 
    # name.tld style.  This is /NOT/ 100% accurate!

    $domain = join '.' =>
      $atom[-2] =~ /^(?:com?|net|org|ac|edu)$/
      ? @atom[-3,-2,-1]
      : @atom[-2,-1];

} else {

    warn("Couldn't extract gTLD/ccTLD from $host\n");
    $domain = undef;

}
[download]

    --k.