3.1.a (i) Address types, VLSM

3.1.a (i) Address types, VLSM


For this section you should understand the basics of IPv4 addressing.

We should remember that an IPv4 address is a 32bit number with a max of about 4 billion.
It’s written in dotted decimal notation like so
Each number separated by a decimal is 1 byte in size.
That means each number has a value possibly from 0 to 255.
Each byte is called an octet, and you may read RFCs where instead of a byte they will say an octet so it’s good to know this term.
There are 8 bits per byte.

Originally we had classes that determined how the networks were broken up. (ABCDE)
The class the address was in was determined by the first 4 bits from the left.

These classes do NOT exist anymore however here are they for reference:

Class A:
In binary starts with a 0
The decimal value is from 0 – 127
8 network bits, 24 host bits

Class B:
In binary starts with a 10
The decimal value for this is 128-191
16 network bits, 16 host bits

Class C:
In binary starts with a 110
The decimal value for this is 192-223
24 network bits, 8 host bits

Class D:
In binary starts with 1110
The decimal value for this is 224-239

Class E:
Starts with binary 1111
The decimal value for this is 240-255

ABC are unicast addresses, D is for multicast, and E was for special / reserved.

The issue with this assignment was that larger networks didn’t actually need a class A or a class B for public addresses.

Then we came up with classful subnetting.

Subnetting allowed an administrator to further divide their network block by borrowing bits from the host field and moving them to the network field.

2 Addresses are lost for each time a subnet is created ( for the network and broadcast)
The exception to this rule is using a /31 which loses only 1 address, the network address.

Originally RIP and IGRP did not support sending the subnet mask over their updates.

The downside of classful subnetting is the subnets we made needed to be made of the same size.

Then we came up with variably length subnet masks (VLSM).

This meant we could have subnets from any length. This required the use of classless routing protocols (RIPv2, ISIS, BGP, OSP, basically anything newer).