Why IPv6’s Idealistic Vision Collides with Real-World Network Constraints
The article examines how IPv6 was designed to eliminate decades‑old networking baggage, but its reliance on discarding legacy layers, MAC addresses, ARP, and DHCP makes the ideal blueprint unattainable in today’s mixed‑technology Internet.
Idealism vs. Reality in IPv6
IPv6 was intended to solve not only address exhaustion but also the accumulated historical baggage of the network stack.
From Point‑to‑Point to Bus Networks
Early networks were point‑to‑point links without addressing. LANs introduced shared media (bus topology) and required Layer 2 addresses. Early Arcnet used 8‑bit manually set addresses; Ethernet adopted globally unique 48‑bit MAC addresses, eliminating collisions but creating long‑term constraints.
Layer 2 and Layer 3 Entanglement
When an IP packet is transmitted over Ethernet, the OS must map the next‑hop IP address to a MAC address using ARP, which broadcasts a query to all hosts. Broadcasts can cause traffic storms, especially on Wi‑Fi. DHCP, effectively a reverse ARP, also ties IP allocation to MAC addresses.
layers are only ever added, never removed.
The Ghost of the Bus Network
Modern Ethernet uses a star topology with switches, and Wi‑Fi operates in infrastructure mode, also a star. Wi‑Fi frames contain three‑address, four‑address, or six‑address modes to simulate point‑to‑point communication over a non‑existent bus.
IPv6’s Blueprint
No physical bus network
No Layer 2 interconnects (Layer 3 should handle it)
No broadcast (use multicast)
No MAC addresses (point‑to‑point links have known endpoints)
No ARP or DHCP
Simplified IP header for hardware acceleration
Ample address space eliminates shortage
If realized, Wi‑Fi access points and Ethernet switches would act as IPv6 routers, eliminating ARP storms. Ethernet frames could drop 12 bytes (source/destination MAC) and Wi‑Fi frames 18 bytes; IPv6 adds a 24‑byte address, resulting in a net increase of only 12 bytes.
Why the Vision Is Unattainable
IPv6 assumes discarding all legacy mechanisms, but IPv4, MAC addresses, and existing frame formats persist. Consequently IPv6 still requires Neighbor Discovery, simulated broadcast, DHCP for legacy devices, and NAT for many IoT gadgets.
Mobile IP: The Forgotten Killer App
IPv6’s original design omitted Mobile IP. Current seamless handover relies on Layer 2 bridging, creating large two‑layer domains and added latency.
TCP/UDP identify a session with a four‑tuple (src IP, src port, dst IP, dst port). When a device’s IP changes, the tuple changes and connections break.
A pure Layer 4 identifier such as a 128‑ or 256‑bit UUID could allow routers to forward based on IP while the kernel matches packets to sockets by UUID, enabling transparent mobility.
QUIC already uses a unique Session ID, supporting connection roaming and effectively providing a transparent mobile IP.
Practical Takeaways
Perfect design requires a clean slate. IPv6’s theoretical elegance depends on worldwide abandonment of legacy systems, which is infeasible.
Incremental improvement beats a clean‑room rewrite. QUIC adds a session layer on top of UDP without discarding TCP/UDP, solving core problems while staying compatible.
More layers increase inertia. Each protocol layer carries its own hardware, software, operational knowledge, and vested interests, making removal costly.
IPv6 is not a failed design—it is a perfect solution for a world that never existed.
References
Avery Pennarun, “The world in which IPv6 was a good design”, 2017
RFC 1710: IPv4 with more address bits
QUIC Protocol (IETF RFC 9000)
MinimaLT: Minimal‑latency Networking Through Better Security
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.