This approach is very similar to the one that we used before, but instead of storing
IP addresses as two integers, they are encoded as a big-endian 128 bit long binary
"string".
This is then stored as the BLOB data type of sqlite which has fewer limitations in
length, can be indexed and uses memcmp() which should be at least as fast as comparing
integers.
This allows a slightly easier query and seems to me significantly faster than the
PostgreSQL approach.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>