What Is IPv6 and Why Does It Actually Matter

IPv4 gives us about 4.3 billion addresses. The internet now has more than 15 billion connected devices. Do the maths. IPv6 was built to fix this — it offers 340 undecillion addresses, which is a number so large it's basically meaningless to say out loud.

The Address Format Problem

IPv4 addresses look like 192.168.1.1 — four numbers from 0 to 255, separated by dots. Easy enough. IPv6 addresses look like 2001:0db8:85a3:0000:0000:8a2e:0370:7334 — eight groups of four hexadecimal characters, separated by colons. Humans don't read these. Computers don't care. The format is the reason IPv6 adoption dragged — every piece of software, every UI, every log parser had to be updated to handle the new format.

There are shortening rules. Consecutive groups of zeros can be collapsed to ::, so 2001:0db8:0000:0000:0000:0000:0000:0001 becomes 2001:db8::1. Still not pretty, but workable.

Why It's Not Just About More Addresses

IPv6 was redesigned from scratch, not just extended. NAT — the network address translation hack that lets thousands of devices share one IPv4 address — becomes largely unnecessary. Every device gets a globally routable address. This simplifies routing, improves end-to-end connectivity, and eliminates a whole class of NAT traversal problems that plague real-time applications like VoIP and video calls.

IPSec — the protocol suite for encrypted, authenticated IP traffic — is mandatory in IPv6, not optional. Auto-configuration lets devices generate their own addresses without a DHCP server. The header format is cleaner, which means faster routing at the hardware level.

Where IPv6 Actually Is Right Now

Google reports IPv6 adoption at around 45% of its global traffic as of early 2026. Mobile networks were early adopters — T-Mobile US runs almost entirely on IPv6. Residential ISPs are mixed. Business networks are often behind, partly because network engineers are comfortable with IPv4 and partly because the migration requires real work.

Honestly, though — the real barrier isn't technical, it's organisational. Every migration project hits the same problem: someone's legacy application doesn't handle IPv6 addresses correctly, and fixing it requires touching code nobody wants to touch.

Dual-Stack: How the Transition Actually Works

Most modern networks run dual-stack — both IPv4 and IPv6 simultaneously. When you connect to a server that supports both, your device prefers IPv6 (this is called Happy Eyeballs in RFC 8305). If IPv6 fails, it falls back to IPv4 without interruption. You probably won't notice the difference. But your traffic is increasingly taking the IPv6 path without you knowing it.

IPv6 Privacy Addresses

IPv6 introduces a privacy concern that IPv4 didn't have. In IPv4, your device's address is assigned by DHCP and changes. In early IPv6, the global part of your address was derived from your network interface's MAC address — which is hardware-fixed and permanent. That meant your IPv6 address was a globally unique, persistent identifier that followed your device across every network it joined.

RFC 4941 addresses this with privacy extensions — generating temporary, randomised IPv6 addresses that change periodically (typically every 24 hours). All modern operating systems implement this. But it's worth knowing the setting exists and confirming it's enabled if IPv6 privacy matters to you. On Windows, run 'netsh interface ipv6 show privacy' to check. On Linux, check the use_tempaddr setting in the kernel network parameters.