An internet protocol (IP) address is a way of telling every other device on a network exactly where a particular device can be reached over an IP network. In some ways, it's similar to the way your mail delivery person uses your house address to deliver your mail, except that computer data sent to you is never delivered to your neighbor's computer by mistake, and an IP network doesn't take the day off for holidays.
Like your street address, which has a street number and name and a postal code, an IP address has a couple of parts. There's a network part and a "host" part. Think of the network part of an IP address as the postal code. It routes data to the right general area but not straight to your door. Then there's the host part, which identifies the exact location of the host or the computer to which the data is intended.
Unlike a postal code, the network portion of an IP address is variable in length, depending on the size of the network. IP addresses are currently formatted in what's called "quad-dotted decimal notation." That's a geeky way of saying it has four segments, each of which are separated by a dot, and although a computer sees them as an eight-bit binary number, we notate it in decimal because it's much easier for humans to read. So IP addresses range from 0.0.0.0 to 256.256.256.256. If the network is to accommodate, say, 32 computers, then the host portion will need to be able to represent 32 unique numbers in binary. That takes five bits because 00000 in binary is 0 in decimal and 11111 in binary is 31 in decimal. Since the entire IP address is 32 bits long, that leaves 27 bits for the network portion of the IP address.
From "Focus on Fundamentals" by Richard Cadena, PLSN, Dec. 2009
