IPv4 vs IPv6: Differences, Formats and What It Means for Your Services.

IPv4 has carried the internet since the early 1980s, and its roughly 4.3 billion addresses ran out years ago. IPv6 was designed to replace it, but the two will run side by side for a long time. If you run websites or servers, it pays to know where they differ and what can break.

Key takeaways

• IPv4 addresses are 32 bits (203.0.113.10); IPv6 addresses are 128 bits (2001:db8::10).
• IPv6 removes the need for NAT and lets every device have a public address.
• Hosts usually configure IPv6 addresses automatically (SLAAC) as well as via DHCPv6.
• Firewalls must be configured for both protocols - an open IPv6 path is a common oversight.
• If you publish an AAAA record, monitor the IPv6 path separately.

Address format

Shortening IPv6 addresses

Two rules make IPv6 addresses readable:

1. Leading zeros in a group can be dropped: 0db8 becomes db8.
2. One run of all-zero groups can be replaced with ::.

2001:0db8:0000:0000:0000:0000:0000:0010
2001:db8::10

In URLs, IPv6 addresses go in square brackets: https://[2001:db8::10]:8443/.

Main differences

No address shortage, no NAT

IPv4 networks hide many private devices behind one public address using NAT. With IPv6 every device can have a globally unique address, which simplifies peer-to-peer connections - and means the firewall, not NAT, is what keeps inbound traffic out.

Automatic configuration

IPv6 hosts can build their own addresses from the network prefix announced by the router (SLAAC). DHCPv6 is also available when you need central control.

Simpler header, no fragmentation by routers

The IPv6 header has a fixed size, and routers never fragment packets; senders discover the path MTU themselves. Firewalls that block all ICMPv6 break this discovery, causing connections that hang on large responses.

ICMPv6 is essential

IPv6 relies on ICMPv6 for neighbour discovery and path MTU. Blocking it wholesale - a common IPv4 habit - breaks things. Allow the required types instead.

Running services on both (dual stack)

Most servers today run both protocols. Clients that support IPv6 usually try it first and fall back to IPv4 ("happy eyeballs"). That creates a failure mode that is easy to miss: the site works for you over IPv4 while some users reach a broken IPv6 path.

Checklist for dual-stack services:

1. The web server listens on IPv6 too (listen [::]:443; in nginx).
2. The firewall has matching IPv6 rules - neither too closed nor wide open.
3. The TLS certificate and virtual hosts work identically on both.
4. The AAAA record points to the right address, or doesn't exist if IPv6 isn't ready.
5. Logs and rate limits handle IPv6 addresses (and /64 prefixes) correctly.

Monitoring IPv4 and IPv6

If you publish both an A and an AAAA record, check both paths. A monitor that resolves the hostname may pick either protocol, so set up separate checks against the IPv4 and IPv6 address, or a DNS monitor that confirms both records hold the expected values.

Frequently asked questions

• Is IPv6 faster than IPv4?

  +

  Not inherently. Performance depends on the route each protocol takes, which can differ. Sometimes IPv6 is faster, sometimes slower.

• Do I have to support IPv6?

  +

  Not yet everywhere, but more mobile and home networks prefer it each year. If you do publish an AAAA record, make sure it works - a broken IPv6 path is worse than none.

• Can IPv4 and IPv6 hosts talk to each other directly?

  +

  No. They are separate protocols; translation mechanisms such as NAT64 are needed between an IPv6-only and an IPv4-only host.