Routing simplification
To simplify routing issues, IPv6 addresses are considered in two parts: a prefix and an ID. This might seem the same as the IPv4 net-host address breakdown, but it has two advantages.
| Item | Description |
|---|---|
| no class | No fixed number of bits for prefix or ID, which allows for a reduction in loss due to over-allocation |
| nesting | An arbitrary number of divisions can be employed by considering different numbers of bits as the prefix. |
Case 1
| 128 bits |
|---|
| node address |
Case 2
| Item | Description |
|---|---|
n bits |
128-n bits |
Subnet prefix |
Interface ID |
Case 3:
| Item | Description | |
|---|---|---|
n bits |
80-n bits |
48 bits |
Subscriber prefix |
Subnet ID |
Interface ID |
Case 4:
| Item | Description | ||
|---|---|---|---|
s bits |
n bits |
m bits |
128-s-n-m bits |
Subscribe prefix |
Area ID |
Subnet ID |
Interface ID |
Generally, IPv4 cannot go beyond Case 3, even with Variable Length Subnet Mask (VLSM is a means of allocating IP addressing resources to subnets according to their individual need rather than some general network-wide rule). This is as much an artifact of the shorter address length as the definition of variable length prefixes, but is worth noting nonetheless.