IPv4 vs IPv6: What's the Difference?
IPv4 uses 32-bit addresses (about 4.3 billion total, written like 192.0.2.1),
while IPv6 uses 128-bit addresses (about 3.4×1038 total, written like
2001:db8::1). The core reason IPv6 exists is that the world ran out of free IPv4
addresses; IPv6 also removes the need for NAT and restores true end-to-end addressing.
IPv4 vs IPv6 comparison table
| Feature | IPv4 | IPv6 |
|---|---|---|
| Address length | 32 bits | 128 bits |
| Total addresses | ~4.29 billion (232) | ~3.4×1038 (2128) |
| Notation | Dotted decimal, e.g. 192.0.2.1 |
Hexadecimal, colon-separated, e.g. 2001:db8::1 |
| Header size | 20–60 bytes (variable, optional fields) | Fixed 40 bytes (extension headers chain) |
| Header checksum | Yes (recomputed at each hop) | Removed (relies on link/transport layers) |
| Address configuration | Manual or DHCP | SLAAC (auto), DHCPv6, or manual |
| NAT | Ubiquitous (to conserve addresses) | Not needed; end-to-end addressing |
| Broadcast vs multicast | Uses broadcast + multicast | No broadcast; multicast and anycast only |
| IPsec | Optional add-on | Designed in (still optional in practice) |
| Address-to-MAC resolution | ARP | Neighbor Discovery (NDP, ICMPv6) |
Address space: why the size matters
IPv4's 32-bit address space tops out at 4,294,967,296 addresses. That sounded limitless in the early 1980s, but a planet with billions of phones, laptops, servers, IoT sensors and cloud instances exhausted it. IANA allocated its last free IPv4 blocks to the Regional Internet Registries in February 2011, and most RIRs have since reached effective depletion, forcing the rise of a secondary market where IPv4 blocks trade for real money.
IPv6's 128-bit space is not "four times larger" - it is exponentially larger. The
difference (2128 vs 232) is a factor of 296, roughly 79
octillion. A single residential customer is commonly delegated a /64 subnet,
which alone holds 18,446,744,073,709,551,616 addresses. This abundance is what lets IPv6
drop NAT and give every device its own routable address.
Notation and shorthand rules
IPv4 addresses are four 8-bit octets in decimal, separated by dots:
203.0.113.42. IPv6 addresses are eight 16-bit groups in hexadecimal, separated
by colons: 2001:0db8:0000:0000:0000:ff00:0042:8329. Two compression rules make
IPv6 readable:
- Leading zeros in a group may be dropped:
0db8→db8,0042→42. - One run of consecutive all-zero groups may be replaced with a double colon
::(used at most once per address). The example above shortens to2001:db8::ff00:42:8329.
The loopback is 127.0.0.1 in IPv4 and ::1 in IPv6. The
documentation ranges reserved by RFC are 192.0.2.0/24,
198.51.100.0/24, 203.0.113.0/24 for IPv4 and
2001:db8::/32 for IPv6.
Adoption and the dual-stack transition
There is no single switch-over. The internet runs dual-stack: hosts and routers speak both protocols, preferring IPv6 when both a route and a AAAA record exist (per the Happy Eyeballs algorithm, RFC 8305, which races connections to avoid stalls). Google's public measurements have shown IPv6 reaching users in the 40–50% range globally, with some countries well above that and others still low, so IPv4 remains essential for universal reachability.
Because the two protocols cannot talk directly, mixed environments rely on transition mechanisms: NAT64/DNS64 and 464XLAT let IPv6-only clients reach IPv4-only servers, while tunneling carries one protocol inside the other across legacy links.
Related tools
- IPv4 to IPv6 Converter - map an IPv4 address into IPv6 notation.
- IP Lookup - geolocation and network details for any IPv4 or IPv6 address.
- IP Validation - check whether a string is a valid IPv4 or IPv6 address.
- IP Subnet Calculator - work out subnet ranges and host counts.
Frequently asked questions
Is IPv6 faster than IPv4?
Not inherently. IPv6 removes per-packet header checksums and avoids NAT, which can reduce processing overhead, but raw throughput depends on your network path, hardware and ISP. In practice the speed difference for most users is negligible; the real benefit of IPv6 is the vastly larger address space and end-to-end addressing.
Will IPv4 be turned off?
No fixed shutdown date exists. The internet is running a long dual-stack transition where devices speak both protocols at once. IPv4 will remain reachable for years because so much infrastructure depends on it, but new address allocations are scarce and expensive, which is why adoption of IPv6 keeps climbing.
How many addresses does IPv6 actually provide?
IPv6 uses 128-bit addresses, giving 2^128 (about 3.4 x 10^38) possible addresses. IPv4 uses 32-bit addresses for 2^32 (about 4.29 billion). The difference is so large that a standard home connection is typically handed an entire /64 block - 18.4 quintillion addresses - for one customer.
Do I still need NAT with IPv6?
Generally no. IPv6 has enough addresses to give every device a globally routable address, so the address-conservation reason for NAT disappears. Networks still use a stateful firewall for security, and some deployments use NPTv6 for prefix translation, but classic many-to-one NAT (NAPT) is no longer required.
Can IPv4 and IPv6 talk to each other directly?
No. The two protocols are not wire-compatible, so an IPv4-only host cannot directly reach an IPv6-only host. Connectivity between them requires a translation or tunneling mechanism such as NAT64/DNS64, 464XLAT, or dual-stack hosts that run both protocols simultaneously.
How this tool works: This tool runs in your browser and on our server in real time. Depending on the tool, results are computed directly from the input you provide or retrieved from live, authoritative data sources at the moment you run a lookup. We do not sell your data, and your lookups are kept private — any history shown here is stored only on your device.