Page 159 - DCAP207_NETWORKS_DCAP406_COMPUTER_NETWORKS

Computer Networks/Networks

Notes 10.1.3 Subnetting for IP Addresses

Over the past several years, the Internet has scaled enormous volume in terms of hosts connected
to it and therefore IPv4 addresses yet available are becoming scare. You may have confusion
here that 32 bits give 2 unique addresses which comes around 4.3 billion different addresses.
32
But this not the condition because of the different classes of the IPv4 addresses. Suppose a
medium sized organization gets Class B address based on its current user population of say
1000. It uses 1000 different addresses. But the organization management has the ability to assign
16
2 = 65,536 different identifiers. It means that there is 65,536 addresses wastage. Since they all
belong to the same class B network number, they cannot be reclaimed by any other organization.
A network administrator may suggest using Class C network address, which may require at
least four class C networks. Later on, suppose, the number of users increase and the organization
applies for another class C network, it might not get the same or if it gets, it has to pass through
a hell of paper works and delays. In addition there is another angle of this problem with regard
to additional routing. With many Class C networks, you need to have more network number
for routers to track. Consequently performance of the network deteriorates. The solution of
these problems lie either in increasing the number of bits in IP address or Classless Inter
Domain Routing (CIDR).

We may also use a technique called subnetting to efficiently divide the address space allocated
to an organization to the different users divided among different subnets of an organization
network. Therefore subnetting is a process through which the address space of a unicast address
prefix is efficiently divided for allocation among the subnets of an organization network. As we
know that a unicast address have fixed and variable portions. The fixed portion of a unicast
address prefix has a defined value. The variable portion of a unicast address prefix has the bits
beyond the prefix length, which needs to set to 0. Subnetting uses the variable portion of a
unicast address prefix for assignment to the subnets of an organization network.
In order to implement subnetting, you need to follow the some guidelines:

 Assess the number of subnets requirement.
 Assess the number of host IDs for each subnet.
After this, a set of subnetted address prefixes with a range of valid IP addresses may be defined.
Following steps are followed for subnetting:

1. Estimate the number host bits for the subnetting.
2. Determine the new subnetted address prefixes.
3. Determine the range of IP addresses for each new subnetted address prefix.

We may now learn as to how the subnet prefix of an IP address is determined. Following steps
give you a way to determine the same without the use of binary numbers:
1. Write the number n (the prefix length) as the sum of 4 numbers by successively subtracting
8 from n. For example, 22 is 8 + 8 + 6 + 0.
2. In a table with four columns and three rows, place the decimal octets of the IP address in
the first row. The second row will then contain the four digits of the sum as has been
determined in step 1.
3. The columns having 8 in the second row, write the corresponding octet from the first row
to the third row. In case of 0 in a column in the second row, place 0 in the third row.
4. The column in the second row having a number between 0 and 8, convert the decimal
number in the first row to binary. Now select the high-order bits for the number of bits

152 LOVELY PROFESSIONAL UNIVERSITY