User manual

Addressing Schemes B-

Subnetting

A large IP network can be subdivided into smaller subnetworks.

This is done using a device called the subnet mask (in this text,

often called netmask), which tells a routing device how to

further subdivide the Host ID portion of an IP address.

A subnet mask is a 32 bit value which is written in dotted

decimal notation. It contains a number of bits set to 1

(indicating the network portion of an address) followed by a

number of bits set to 0 (indicating the host portion of an

address).

For example, a netmask of 255.255.255.0 on a Class B network

would indicate that the network is divided into 254

subnetworks of 254 nodes each (0 and 255 are reserved

numbers). 128.5.63.28 would be host 28 on subnetwork 63 of

that network. The natural network itself would be called

128.5.0.0 (Class B network number 5).

Notice that by using subnet masks, you can define a natural

hierarchy in which the addresses themselves indicate how a

packet is to be routed. However, all routing devices on an IP

network must be using the same subnetting scheme.

Also note that a subnet mask for a given network segment is not

part of the address and is not transmitted with every packet. It

is simply a value which is known to all the routing devices

adjacent to that segment.