I concur with salva. Once you expand your ranges and store them sorted in a packed binary file, lookup via binary search are very fast.

I generated a file of 125,000 ips (an average of 125 allocated to each of 1000 users).

Single lookups with a cold cache take about 1/100th of a second, which is probably quicker than you could load the file of ranges and populate a hash:

C:\test>echo 92.103.2.204 | 874413.pl
92.103.2.204 allocated to user;552
1 lookups (1 found) in 0.011 seconds (0.011339/sec)

C:\test>echo 63.203.110.110 | 874413.pl
63.203.110.110 allocated to user;532
1 lookups (1 found) in 0.012 seconds (0.012047/sec)

C:\test>echo 109.224.107.185 | 874413.pl
109.224.107.185 allocated to user;139
1 lookups (1 found) in 0.012 seconds (0.012018/sec)

C:\test>echo 122.228.116.36 | 874413.pl
122.228.116.36 allocated to user;485
1 lookups (1 found) in 0.013 seconds (0.012670/sec)

C:\test>echo 0.0.0.0 | 874413.pl
1 lookups (0 found) in 0.007 seconds (0.007276/sec)

C:\test>echo 127.0.0.1 | 874413.pl
1 lookups (0 found) in 0.007 seconds (0.007116/sec)

C:\test>echo 2.103.2.204 | 874413.pl
1 lookups (0 found) in 0.008 seconds (0.008237/sec)
[download]

Once the cache is hot--as would be the case if doing a large number of lookups--that speeds up considerably to well under 1 millisecond per:

>perl -MList::Util=shuffle 
-nE"chomp(@i=shuffle<>); say +(split':')[0] for @i[0..1e3];last" ips.t
+xt 
|874413.pl >nul
1001 lookups (1001 found) in 0.456 seconds (0.000455/sec)
[download]

My crude test code:

#! perl -slw
use strict;
use Time::HiRes qw[ time ];

open IPS, '<raw', 'ips.bin' or die $!;
my $size = -s( *IPS );
$size /= 6;

sub find {
    local $/ = \6;
    my $toFind = unpack 'N', pack 'C4', split '\.', shift;
    my( $lo, $hi ) = ( 0, $size );
    my( $ip, $u );

    while( $lo <= $hi ) {
        my $pos = ( $lo + $hi ) >> 1;

        seek IPS, $pos*6, 0;
        ( $ip, $u ) = unpack 'Nn', <IPS>;

        $hi = $pos - 1, next if $ip > $toFind;
        $lo = $pos + 1, next if $ip < $toFind;
        return join('.',unpack'C4',pack'N',$ip), $u if $ip == $toFind;
    }
    return;
}

my $start = time;
my $found = 0;
while( <> ) {
    chomp;
    my( $ip, $user ) = find( $_ );
    #warn "$_ unassigned\n" and
    next unless $user;
    ++$found;
    printf "%s allocated to user;%u\n", $ip, $user;
}

my $taken = time() - $start;
warn sprintf "%d lookups ($found found) in %.3f seconds (%.6f/sec)\n",
+ 
    $., $taken, $taken / $.;
[download]