IPv6 subnet splitting follows the same mathematical principle as IPv4, but with 128-bit addresses instead of 32-bit.
Splitting Formula: Number of subnets = 2^(new prefix - old prefix)
Common IPv6 Splitting Scenarios:
- Split /48 into /56 → 2^(56-48) = 256 subnets (recommended for small sites)
- Split /56 into /64 → 2^(64-56) = 256 subnets (each /64 for a VLAN)
- Split /64 into /66 → 2^(66-64) = 4 subnets (unusual, /64 is normally not split)
- Split /32 into /48 → 2^(48-32) = 65,536 site allocations
IPv6 Best Practices:
- /64 is the standard subnet size for end networks (SLAAC requirement)
- /48 is the typical allocation for a site or customer
- /56 is common for small sites or home networks
- Splitting on nibble boundaries (/4, /8, /12, etc.) makes addresses more readable
Example: Splitting 2001:db8::/48 into /56 creates:
- 2001:db8:0::/56, 2001:db8:1::/56, ... 2001:db8:ff::/56 (256 subnets)
Unlike IPv4, IPv6's abundant address space means you rarely need to conserve addresses, making generous subnet allocations standard practice.