How does the binary DAT from Maxmind work?

Question

Maxmind offers a binary DAT file format for downloading their GeoIP database.

http://geolite.maxmind.com/download/geoip/database/GeoLiteCity.dat.gz

Does anyone know how this has been packaged? Also, is there any kind of copy protection on the data?

I'd like to offer up a set of data in a similar way.

Anyone with any knowledge of this will receive my undying gratitude :-)

Answer 1

It's just a proprietary binary format, heavily optimized for IP address querying. It doesn't have any copy protection.

If you really want to reverse-engineer the format, take a look at the C# or Java API.

Answer 2

I don't know if this helps but here it's a sample code in PHP for the GeoLite Country DB:

const COUNTRY_BEGIN = 16776960;
const COUNTRY_EDITION = 106;
const STANDARD_RECORD_LENGTH = 3;

public function Seek_Country($ip)
{
    $result = false;

    $databases = glob('./application/repository/GeoIP/GeoIP_*.dat');

    if (array_key_exists(0, $databases))
    {
     rsort($databases);

     if (!$handle = fopen($databases[0], 'rb'))
     {
      return $result;
     }

     $offset = 0;

     flock($handle, LOCK_SH);

     for ($depth = 31; $depth >= 0; --$depth)
     {
      fseek($handle, 2 * self::STANDARD_RECORD_LENGTH * $offset, SEEK_SET);

      $buffer = fread($handle, 2 * self::STANDARD_RECORD_LENGTH);

      $x = array(0, 0);

      for ($i = 0; $i < 2; ++$i)
      {
       for ($j = 0; $j < self::STANDARD_RECORD_LENGTH; ++$j)
       {
        $x[$i] += ord($buffer[self::STANDARD_RECORD_LENGTH * $i + $j]) << ($j * 8);
       }
      }

      if ($ip & (1 << $depth))
      {
       if ($x[1] >= self::COUNTRY_BEGIN)
       {
        $result = $x[1];

        break;
       }

       $offset = $x[1];
      }

      else
      {
       if ($x[0] >= self::COUNTRY_BEGIN)
       {
        $result = $x[0];

        break;
       }

       $offset = $x[0];
      }
     }

     flock($handle, LOCK_UN);
     fclose($handle);
    }

    return $result;
}