IPv4 vs IPv6, explained

Two address formats

IPv4 addresses are 32 bits, written as four decimal numbers separated by dots, such as 203.0.113.24. That format allows roughly 4.3 billion addresses in total — a number that seemed generous in the early 1980s and ran out in practice during the 2010s.

IPv6 addresses are 128 bits, written as groups of hexadecimal digits separated by colons, such as 2001:db8:85a3::8a2e:370:7334. The address space is vast enough that scarcity is no longer the constraint it was. Runs of zeroes are compressed with a double colon, which is why IPv6 addresses vary so much in length on screen.

Why IPv6 exists

IPv6 was standardised primarily to solve address exhaustion. Its current specification is RFC 8200, with the addressing architecture in RFC 4291. Along the way it simplified the packet header, made address autoconfiguration a first-class feature, and removed the need for the widespread address sharing that IPv4 now depends on.

The transition has been slow because the two protocols are not directly interoperable. A device with only an IPv6 address cannot talk to an IPv4-only server without a translation mechanism in between, so the practical answer has been to run both.

Dual-stack connections

Running both protocols at once is called dual-stack. Your connection holds an IPv4 address and one or more IPv6 addresses, and each connection uses whichever protocol both ends support. Browsers generally prefer IPv6 when it is available and fall back quickly when it is not — the behaviour described in RFC 8305.

This is why a tool may report two different public addresses for you at the same moment. They are not in conflict: they are the same connection seen through two protocols.

• IPv4-only: one dotted address, almost always shared through NAT.
• IPv6-only: increasingly common on mobile networks, usually paired with a translation service for IPv4 destinations.
• Dual-stack: both addresses present; which one a site records depends on how the connection was made.

IPv6 is not automatically more private

It is tempting to assume a newer protocol is a more private one. It is not. IPv6 removes the need for address sharing, and address sharing was incidentally providing a small amount of cover: under IPv4 carrier NAT, many subscribers look alike from the outside. Give every device its own globally routable address and that blending disappears.

The mitigations are real but partial. Temporary addresses, specified in RFC 8981 and enabled by default on most modern systems, rotate the host portion of an IPv6 address so it is not derived from a fixed hardware identifier. The network prefix, however, still identifies your network.

Security is a separate question again. IPv6 does not encrypt traffic and does not by itself make a device harder to reach. A device with a globally routable address is directly addressable in a way an address behind NAT is not, so firewall configuration matters more rather than less.

What to check on your own connection

Look at which address a site actually sees, which operator it belongs to, and whether anything is reachable on it. Those facts explain most of what the Internet can currently tell about your connection.