Question : Subnetting

I understand that computers communicate via IP adresses, but what is the purpose of subnetting and how does it work?

Peter

Answer : Subnetting

Phone numbers have structure.  There's a local part and an "exchange" prefix; if you're dialing further away, there's an area code, and even a country code in front of that.

Network addresses have a similar structure:  There's a local (host) part, like the last four digits of your phone number, and a prefix (network) part, like the country code, area code, and exchange.

The analogy starts to break down a bit there, in that the structure of the phone system is fairly rigid.  (But note that if you start dialing with a different country code, the rules for areas and exchanges may not be what you're used to!)

An IP address is 32 bits.  The dividing line between network address part and host address part could be drawn just about anywhere in those 32 bits, and the subnet mask tells us where it's drawn for the local network we're on.

There are contexts in which it's convenient to treat it as being drawn in a different place.  If we're a long-distance carrier trying to deliver a phone call, we might not care what exchange the destination number is in, so long as we deliver the call to a connection point in the right area code.  For our purposes, area codes are the networks we route between, and the exchange codes are subnets that the local phone companies may (or may not!) care about.

Think of "subnets" as networks that can be grouped together because they share a common prefix.  The interconnections that make the Internet work have the option of simplifying their load by treating an entire such group by a single rule, instead of needing to know about every local subdivision of addresses in the world.