The Two Versions of the Internet Protocol
IPv4 and IPv6 are both versions of the Internet Protocol — the foundational set of rules that governs how data packets are addressed and routed across networks. IPv4 has been in use since 1983. IPv6 was standardized in 1998 to address IPv4's limitations, most critically its finite address space. Despite decades of coexistence, the transition is still very much in progress.
Side-by-Side Comparison
| Feature | IPv4 | IPv6 |
|---|---|---|
| Address length | 32 bits | 128 bits |
| Address format | Dotted decimal (e.g., 192.0.2.1) | Colon-hex (e.g., 2001:db8::1) |
| Total addresses | ~4.3 billion | ~340 undecillion |
| Address configuration | DHCP or manual | SLAAC, DHCPv6, or manual |
| NAT required? | Yes (typically) | No (by design) |
| Header size | 20–60 bytes (variable) | 40 bytes (fixed) |
| Broadcast | Yes | No (replaced by multicast/anycast) |
| IPsec support | Optional | Mandatory (though not required in use) |
| Fragmentation | Routers and hosts | Sending host only |
| Checksum in header | Yes | No (handled by upper layers) |
Address Space: The Fundamental Difference
IPv4's 32-bit address space yields roughly 4.3 billion addresses. IPv6's 128-bit space yields approximately 3.4 × 10³⁸ addresses — enough to assign billions of addresses to every grain of sand on Earth. This eliminates address exhaustion as a practical concern for the foreseeable future and removes the need for NAT, which IPv6's designers considered an architectural workaround rather than a proper solution.
Header Simplification
IPv6 uses a fixed 40-byte header, compared to IPv4's variable-length header. Options and extensions in IPv6 are handled through separate extension headers that are only processed by the destination (except hop-by-hop options). This dramatically speeds up router processing, as routers no longer need to parse variable-length headers for every packet.
Autoconfiguration: SLAAC
Stateless Address Autoconfiguration (SLAAC) is one of IPv6's most significant practical improvements. A host can generate its own globally routable IPv6 address by combining the network prefix advertised by its router with a locally generated interface identifier — with no DHCP server required. This simplifies network bootstrapping, especially in large or dynamic environments.
The End of Broadcast
IPv4 uses broadcast for functions like ARP (Address Resolution Protocol), flooding a frame to every device on a segment. IPv6 replaces broadcast entirely with multicast and anycast. Neighbor Discovery Protocol (NDP) uses multicast groups so that only relevant nodes process discovery messages, reducing unnecessary interruptions to every device on a segment.
Coexistence: Not Either/Or
For the foreseeable future, both protocols will coexist. Dual-stack deployments run IPv4 and IPv6 simultaneously, with applications preferring IPv6 when available (per RFC 6724 address selection rules). Pure IPv6-only deployments are growing — particularly in mobile networks and cloud environments — but complete IPv4 retirement remains a long-term horizon for most organizations.
Which Should You Focus On?
- New deployments: Design for IPv6-first with IPv4 as a fallback.
- Existing infrastructure: Incrementally enable dual-stack, starting at the edge.
- Legacy systems: Use NAT64/DNS64 to extend their life while the broader network transitions.
Understanding both protocols deeply — their similarities, differences, and interaction — is an essential competency for any modern network engineer.