To decide whether the source and destination hosts are on the same network looks  easy: just make a comparison between  the two   network IDs of the two addresses. For example consider the following two addresses:

Source                  200.100.55.101

Destination            200.100.66.72

 

   These are class C networks, so the source address has a network ID of 200.100.55.101, and the destination has network ID of  200.100.66.72. Thus they are on different networks. Or are they ? One of the consequences of  having no more class A and class B address blocks is that many large corporations can handle their addressing needs only by obtaining multiple blocks of class C address. So it's entirely possible that  the 200.100.55.101 and 200.100.66.72 networks IDs belong to the same company and could therefore be part of the same network! If so, IP should look at only the first two quads(200.100) to determine whether the address on the same network.

   So how does IP know to compare the first one, two, or three quads? by using a subnet mask. A subnet is a subsection of a network that uses related IP addressees. On a class C network, for example, you could define the first 127 addresses to be on one subnet and the second 127 addresses to be on another subnet. On a larger scale, from the point of view of the internet - which you think of as being the network- each class A, B and C networks is a subnet.

   The subnet mask is a 32-bit value that usually expressed in the same dotted-decimal notation used by IP addresses. The purpose of the subnet mask is to let IP separate the network ID (or, as  you saw in preceding example, part of the network ID) from the full IP address and thus determine whether source and destination are on the same network

   This table spells out the default subnet masks for each type of network class.

      When IP applies the subnet mask to an IP address, the part of that mask is all 0's strips off the corresponding section of the address. consider the following example:

   The mask produces the same result, so these two addresses are on the same network. Now consider the example we used earlier. In this case we need to use a nonstandard mask of 255.255.0.0:

How subnet Mask works ?

   The operation of the subnet is a bit more complex. than we have let on. It's actually a two step processes. In the first step, the IP addresses are both compared bit by bit with the subnet mask using a Boolean AND operation- if both are 1, a 1 returned; otherwise, a 0 is returned:

   Now the two results are compared bit by bit a Boolean Exclusive or (xoR) operation if the both bits are 0 or 1 , a 0  is returned; otherwise; a 1 is returned:

   I f the results of the xoR operation is all 0's, the source and destination are on the same network.