Net::Subnet::Count - Count hosts in named subnets

  use Net::Subnet::Count;
  use IP::Address;

  my $counter = new Net::Subnet::Count;

  $counter->add('subnet-00', new IP::Address("10.0.0.0/24"));
  $counter->add('other', @array_of_ip_addresses);
  $counter->add('other', @another_array_of_ip_addresses);

  $counter->cache(10);

  $counter->count(new IP::Address("10.0.3.17"));
  $counter->count(@array_of_ip_addresses);

  $counter->valcount(new IP::Address("10.0.3.17"), 23);
  @array_of_ipaddr_and_values = (new IP::Address("10.0.3.17"), 23,
				new IP::Address("101.0.23.107"), 2);
  $counter->valcount(@array_of_ipaddr_and_values);

  my $r_count = $counter->result;

  foreach my $subnet (keys %{$r_count}) {
      print "Subnet $subnet had ", $r_count->{$subnet}, " visits.\n";
  }

This module implements a symplistic way to match individual IP Addresses to subnets. It can be used to, among other things, help analyze HTTPD logs.

The following methods are implemented.

->new

Creates a new counter. This method can be called passing as argument a hash where the keys are the name of the subnet group and the values are references to arrays of IP::Address objects referencing each specific subnet. This is probably ok for static initializations.

->add

Adds a subnet group. The first parameter is the name of the group being added. If it's a new name, a new entry will be created. Else the given subnets are added to the existing ones, like in the example above.

->count

Verifies if the IP::Addresses are contained in any of the given subnets. If this is the case, the corresponding totals are updated.

->valcount

The same as ->count but the argument is an array consisting of IP::Addresses and value pairs.

->result

Returns a reference to a hash containing the respective totals for each subnet group. The key to the hash is the subnet name given with ->add, the value is how many IP::Address objects have been found to match that subnet group.

->cache

Since in usual applications IP::Addresses from the same subnet will tend to be grouped in clusters like in the case of HTTPD logs some caching is attempted to speed things up. The caching consists in storing the last few entries matched in an LRU list which is checked before going through all the stored subnets.

This can improve response times if tuned sensibly, however consider that every miss will cause every entry in the cache to be checked twice, one in the cache and one in the normal process so it's important to tune the cache.

The default cache size is 5, which can be changed by calling the ->cache method as in the example. The old value of the cache size is returned.

Luis E. Munoz <lem@cantv.net>. Alvaro Carvajal <alvaro@cantv.net> contributed the valcount method.

perl(1), IP::Address(1).