[detaro]

The dots are just for human readability, showing the address byte-by-byte, and don't have any "meaning" anymore. Geographic structure is also only accidental.

Subdivisions nowadays happen nearly everywhere and noted with a a slash and a number after the IP address (e.g. 215.54.87.0/24 is the network containing all addresses where the first 24 bits are the same as in 215.54.87.0, 215.54.87.0/28 would be all the addresses where the first 28 bits are the same etc).

The rules differ a bit between the regions, but if I remember correctly a /24 is the smallest that can be officially assigned to an entity and announced in the global routing tables. They are free to slice smaller inside their networks though, and e.g. give a customer a /28 or /30 to use.

I hope that is somewhat understandable, I'm having a hard time putting it in words...

EDIT: IPv6 is same principle (although the notation for the address itself is now different). One important size there is /64, which is the subnet size for which automatic address assignment works, and the default size for subnets containing end devices. If you get IPv6 internet, you get at least a /64 (although a /56 is recommended), so you don't need NAT because you have more than enough addresses for everything.