The size of an IPv6 address is 128 bits. Because the IPv6 addresses are so large, they are difficult to represent in decimal notation as we do in IPv4. That is why to represent an IPv6 address it was decided to use 8 groups of 16 bits in hexadecimal notation, separating each group by the colon character, :. Here, we will see an example of IPv6 address:

fe80: 0000: 0000: 0000: ac3e: 7bff: fe33: 5fb0

Among the different rules that IPv6 has for representation, one of them is that you can suppress the leading zeros in each group to represent the address in the following way:

fe80: 0: 0: 0: ac3e: 7bff: fe33: 5fb0

In addition to this simplification, there is another one that indicates that several groups of consecutive zeros can be replaced by two colons, ::, and this can only be done once in the complete address. Therefore, the previous example could be summarized in:

fe80 :: ac3e: 7bff: fe33: 5fb0

To determine in IPv6 which part of the IP address corresponds to the network and the host, the address must be divided into two 64-bit parts, leaving the previous IP address as follows:

Network address -> fe80 ::
Host address -> ac3e: 7bff: fe33: 5fb0

It should be noted that, in IPv6, the part of the address that corresponds to the host address is fixed. In link-local, as we reviewed before, the host address corresponds to the Mac address of the interface in EUI-64 format.