IPv6 prefix and subnet calculator
Enter an IPv6 address with its parent prefix and choose the child prefix you want to plan.
What is IPv6 and why does it use 128-bit addresses?
IPv6 is the Internet Protocol addressing architecture based on 128-bit addresses. Unlike IPv4, its design provides an enormous addressing space and supports a hierarchical model in which operators normally plan networks by prefixes rather than by trying to conserve individual host addresses.
IPv6 defines unicast, anycast and multicast addressing. It does not define broadcast addresses; multicast mechanisms perform functions that in IPv4 may rely on broadcast.
How IPv6 prefixes work
The number after the slash identifies how many high-order bits belong to the prefix. A /48 fixes the first 48 bits, a /56 fixes 56, and a /64 fixes 64. Moving from a shorter parent prefix to a longer child prefix borrows bits for subnet identification. The number of resulting child prefixes is 2 raised to the difference between the two prefix lengths.
Example: /48 to /64 uses 16 additional subnet bits, producing 65,536 /64 networks. A /56 to /64 uses 8 bits and produces 256 /64 networks.
Why /64 is the normal IPv6 LAN boundary
The 64-bit boundary is deeply embedded in IPv6 architecture and operational practice. Most conventional IPv6 LANs use a /64, particularly where Stateless Address Autoconfiguration (SLAAC) is expected. This does not mean every IPv6 prefix everywhere must be /64: standards define specific exceptions and operational cases, but /64 should be the default assumption for ordinary host subnets unless there is a documented reason to do otherwise.
IPv6 /127 on point-to-point router links
RFC 6164 standardizes /127 prefixes for inter-router point-to-point links. This is a specific infrastructure use case, comparable in spirit to IPv4 /31 addressing. It applies to links with two routers and no hosts; it should not be generalized as a replacement for /64 on ordinary LANs.
/48, /56 and the myth of one universal end-site prefix
IPv6 end-site assignments are intentionally not reduced to one mandatory size. RFC 6177 moved away from treating /48 as a universal default requirement. The operational goal is to assign enough address space for a site to create an appropriate number of subnets without forcing IPv4-style address conservation. Provider and regional Internet registry policies determine the actual allocation or assignment available to a customer.
Global Unicast, ULA and link-local addresses
Global Unicast addresses are intended for normal globally routed IPv6 communication. Unique Local Addresses use FC00::/7, with locally generated ULAs normally using the locally assigned format described by RFC 4193; they are intended for local communications and are not expected to be routed on the global Internet. Link-local addresses use FE80::/10 and are fundamental to communication on the local link, including core IPv6 neighbor and router functions.
Unicast, anycast and multicast — and no IPv6 broadcast
Unicast identifies one interface. Anycast uses an address from unicast space on multiple interfaces and routing delivers traffic to one of them, normally the nearest according to routing metrics. Multicast identifies a group of interfaces. IPv6 has no broadcast address type.
SLAAC, interface identifiers and temporary addresses
SLAAC allows hosts to configure IPv6 addresses using information advertised by routers. Modern IPv6 should not be understood as simply embedding a device MAC address forever in every address. Systems can use stable identifiers and can also generate temporary addresses with randomized interface identifiers to reduce long-term address-based activity correlation.
Compressed and canonical IPv6 notation
IPv6 addresses are written as eight 16-bit hexadecimal fields. Leading zeros can be suppressed and one longest sequence of zero fields can be compressed with ::. RFC 5952 recommends a consistent canonical text representation, including lowercase hexadecimal and suppression of unnecessary leading zeros. The calculator accepts valid compressed input and shows normalized and fully expanded forms.
Use the documentation prefix in examples
The prefix 2001:db8::/32 is reserved for documentation and examples. Using it in tutorials, diagrams and training material avoids accidentally publishing or configuring globally assigned production addresses.
Think in hierarchy, not IPv4-style address conservation
Good IPv6 addressing plans leave room for aggregation, site growth, VLANs, infrastructure, services and future changes. The abundance of addresses is intentional. The main design problem is usually building a clean, summarizable and operationally understandable hierarchy rather than minimizing the number of unused addresses.