Guides · Networking basics

IPv4 vs IPv6: what actually changes for you

Written by Aldo · blue-team threat hunter
Last updated November 9, 2026 · 4 min read

Check your IP on two different days and you may see 203.0.113.45 once and something like 2001:db8:4a2b:7c00:1d3:9f2:aa41:6e05 the next. Same connection, two kinds of address. Here's what the second one changes for you, and what it doesn't.

The short history

An IPv4 address is 32 bits, written as four numbers: 192.0.2.1. That allows about 4.3 billion addresses, and the central pool had run out by 2011. Since then, sharing (NAT at home, CGNAT at providers) has kept IPv4 going.

An IPv6 address is 128 bits, written as eight groups of hexadecimal: 2001:0db8:1234:5678:0000:0000:0000:0001, which may be shortened by dropping leading zeros and replacing one run of zero groups with ::, giving 2001:db8:1234:5678::1. There are enough for every device to have its own globally unique addresses, many times over.

A worked example: one address, many notations

To see how the two systems relate, we put the documentation address 192.0.2.1 into iplover's IP converter:

Two lessons hide in this table. First, an IPv4 address is just a 32-bit number; the dots are for humans. C0 00 02 01 in hex is 192 0 2 1. Second, IPv6 can carry an IPv4 address inside it. You'll meet ::ffff:192.0.2.1 in server logs: it's a plain IPv4 visitor, recorded by software that handles both protocols on one socket. Treat it as 192.0.2.1. (The 6to4 form belongs to an old transition method that's now deprecated; you'll only see it in old documentation.)

What changes for you

Each device has a public address

With IPv4 at home, every device hides behind your router's one address. With IPv6, your provider gives your home a block (commonly a /56 or a /48), your router hands each device addresses from it, and every one is reachable in principle. In practice your router's firewall blocks unsolicited incoming connections by default, which gives you the same protection NAT gave you by accident.

Port forwarding becomes a firewall rule

There's nothing to translate, so there's no forwarding. To make a home server reachable over IPv6, you allow the port for that device's address in the router's IPv6 firewall settings. Behind CGNAT, this is often the only way to accept incoming connections at all.

Your devices change address on purpose

To stop sites from following one device by its address, modern systems use temporary “privacy” addresses for outgoing connections and replace them regularly, typically daily. Windows, macOS, iOS and Android do this by default. So seeing a different IPv6 address tomorrow is normal, but the first half (the network prefix from your provider) usually stays the same for longer, and that half still identifies your connection.

Sites see whichever you used

When both work, your device tries IPv6 first and falls back to IPv4 quickly if it's slow. A site that only has IPv4 sees your IPv4 address; one with both usually sees IPv6. That's why iplover's homepage may show a different kind of address than another checker.

If you see this, do this

What doesn't change

An IPv6 address reveals no more about you than an IPv4 one: the provider, a rough location, and nothing like a home address. It isn't “less private” because it's longer; with temporary addresses it can even change more often. And most of what you do online works identically on both: the difference matters only when you run servers, use VPNs or troubleshoot.