Understanding IPv6 Address Architecture (RFC 4291)
Unlike IPv4's 32-bit limit, IPv6 uses a 128-bit address space, providing approximately 3.4 × 10^38 unique addresses. A standard unicast address is divided into three functional tiers:
|<------------ 64 Bits ------------>|<------ 64 Bits ------>|
[ Global Routing Prefix (usually /48): Assigned by the RIR or ISP to an enterprise site. ] [ Subnet ID (16 bits): Allows an organization to create up to 65,536 individual /64 subnets. ] [ Interface ID (64 bits): The host portion, configured automatically via SLAAC or DHCPv6. ]
- Global Routing Prefix (usually /48): Assigned by the RIR or ISP to an enterprise site.
- Subnet ID (16 bits): Allows an organization to create up to 65,536 individual /64 subnets.
- Interface ID (64 bits): The host portion, configured automatically via SLAAC or DHCPv6.
Why /64 is the Non-Negotiable Standard Subnet
In IPv6, address conservation is unnecessary. RFC 4291 and RFC 4862 establish that all end-user subnets must use a /64 prefix length.
Stateless Address Autoconfiguration (SLAAC) relies on this 64-bit boundary. Subnetting beyond /64 breaks SLAAC and several hardware packet-forwarding optimizations.
Point-to-Point Links: RFC 6164 (/127) vs RFC 3627 (/126)
For router-to-router links, RFC 6164 officially standardizes 127-bit prefixes (/127) to prevent ping-pong packet amplification attacks.