IETF Draft Lets IPv4 Run ARP-Free as a Service on IPv6-Only Nets
New IETF draft removes ARP and IPv4 subnets from IPv6-only networks, turning IPv4 into a pure host service via a sentinel gateway address.
Born from a lightning talk at RIPE 91, the Internet-Draft draft-vanmook-intarea-ipv6-resolved-gateway is now before the IETF IntArea working group, with a presentation set for IETF 126 in Vienna and an adoption call to follow. It aims to strip IPv4 subnets and ARP out of IPv6-only networks without tunnelling or translation.
The motivation is the hidden cost of dual-stack: ARP's operational fragility at scale (storms, cache poisoning, table exhaustion), the address economics of wasted IPv4 space in every subnet, the administrative burden of subnet sprawl, and the security exposure of large ARP-scannable blocks. Hosting providers like Hetzner, OVHcloud, and Scaleway have already been working around this in production for years, each with its own incompatible gateway address and OS-specific configuration hack — an interoperability gap the IETF is now positioned to close.
The fix leaves DHCPv4's Router Option (Option 3) untouched and instead defines a single sentinel address, 192.0.0.11. A host stack aware of the sentinel resolves the gateway's MAC address from its IPv6 neighbor cache instead of sending an ARP request, while unmodified hosts continue to ARP as before — meaning no flag day and fully incremental rollout.
On the network side, the mechanism dovetails with RFC 8950 (IPv4 prefixes with IPv6 next-hops) and the in-progress draft-ietf-intarea-v4-via-v6, closing the host-side gap that draft leaves open. For operators, it turns IPv4 from a pervasive network architecture into what it always was to applications — a service endpoint identifier — carried over IPv6-only transport.