Terms used throughout the pages on Network Setup and Testing are defined here.
IP address: a numerical identification of a particular machine, used to allow other machines wishing to communicate to send data along the correct network segments, so that the data reaches its intended target.
host address: The right-hand portion of an IP address. The boundary between this portion and the left-hand portion (the network-number) is identified by the netmask.
netmask: A constant that defines what part of an IP address is used as the network number, and what part is used as the host address. It is a 32-bit integer, expressed in the same form as an IP address, but with ones in every position corresponding to the network-number field, and zeroes in every position corresponding to the host-number field.
The network-number begins at the most-significant bit position, and extends to the right, ending at an internal boundary. The host address begins at that internal boundary, and extends to the least-significant bit of the 32-bit address. An example will show this more clearly:
|Binary mask||11111111||11111111||11111 000||00000000|
Not all values are legal in a netmask: because the two parts of an IP address must be contiguous, the legal values, with their decimal equivalents, are:
network address: For IPv4 (Internet Protocol version 4), a network address is a 32-bit integer, represented as four 8-bit fields, each expressed as a decimal number eg, 18.104.22.168. Network addresses are separated into two parts for many purposes: the network-number and the host-address. The separation is specified by a netmask.
Much of the confusion about these fields comes up because of the dotted-decimal representation of IP addresses and netmasks. The original formulators thought that hex or octal notation would be too much trouble for users to understand, and so foisted this scheme on an unsuspecting public. They are sorry.
network-number: the left-hand part of an IP address. The network number must be different for each network segment that a host is aware of. (Before Private Networking, each network segment in the world was supposed to have a distinct network number. That is no longer the case, but a particular machine cannot know that there are other Private Networks in the world, so it operates under the blissful supposition that each network number corresponds to a unique network segment.
subnet: The network-number field of an IP address is sometimes further divided into smaller pieces called subnets, by adjusting the netmask. This distinction is not of much use except to operators of large networks; it harks back to when byte-boundaries in the IP address had special meaning.