Operation Manual

IP
SIP Software Release 6-0-0
J613-M0524-00
23
172.16 is the DDN assigned network number
.9 is the user assigned subnet number
.190 is the user assigned host number
The value 0.0.0.0 is used to define the default address, while a value of all
ones in any host portion (i.e. 255) is reserved as the broadcast address. Some
older versions of UNIX use a broadcast value of all zeros, therefore both the
value ‘0’ and the value ‘255’ are reserved within any user assigned host
portion. The address 172.16.0.0 refers to any host (not every host) on any
subnet within the class B address 172.16. Similarly 172.16.9.0 refers to any
host on subnet 9, whereas 172.16.9.255 is a packet addressed to every host on
subnet 9. The router uses this terminology to indicate where packets are to
be sent.
An address with ‘0’ in the host portion refers to ‘this particular host’ while
an address with ‘0’ in the network portion refers to ‘this particular network’.
As mentioned above a value of all ‘1’ (255) is a broadcast. To reduce loading,
IP consciously tries to limit broadcasts to the smallest possible set of hosts,
hence most broadcasts are ‘directed’. For example 172.16.56.255 is a
broadcast to subnet 56 of network 172.16. A major problem with the IP type
of addressing is that it defines connections not hosts. A particular address,
although it is unique, defines a host by its connection to a particular
network. Therefore if the host is moved to another network the address
must also change. The situation is analogous to the postal system. A related
problem can occur when an organization that has a class C address finds
that they need to upgrade to class B. This involves a total change of every
address for all hosts and routers. Thus the addressing system is not scalable.
Subnets
Related to the two issues discussed above, the rapid growth of the Internet
has meant a proliferation in the number of addresses which must be handled
by the core routers. More addresses means more loading and tends to slow
the system down. This is overcome by minimising the number of network
addresses by sharing the same IP prefix (the assigned network number) with
multiple physical networks. Generally these would all be within the same
organisation, although this is not a requirement. There are two main ways of
achieving this; Proxy ARP and subnetting. Proxy ARP will be discussed later
in this section.
A subnet is formed by taking the host portion of the assigned address and
dividing it into two parts. The first part is the ‘set of subnets’ while the
second refers to the hosts on each subnet. For example the DDN may assign
a class B address as 172.16.0.0. The system manager would then assign the
lower two octets in some way which makes sense for this particular
network. A common method for class B is to simply use the higher octet to
refer to the subnet. Thus there are 254 subnets (0 and 255 are reserved) each
with 254 hosts. These subnets need not be physically on the same media.