Use case 03 of 03 plan reversible
One control plane
instead of five.
The transition case is simplification, not renumbering. Fewer moving parts in the core, no per-tunnel state to scale — and per-flow behaviours a full mesh cannot express at all.
5→1
39,800→0
0 bytes
added header, up to six segments An SRH carrying the same six segments adds 104 bytes (8 + 6×16). Six 16-bit instructions after a 32-bit block is exactly 128 bits — the destination address itself · RFC 8754 · RFC 8986What we move: the core stops carrying per-flow state it never needed. No MTU erosion, no hardware header-parsing requirement, and a plan that strands nothing on the way there.
0100 why it collapses
The destination stopped being a place. It became a program.
Six hops fit in the destination address itself. Each node matching the leading instruction shifts the address left and forwards on what is now in front of it — no added header, no per-flow state anywhere in the middle.
The honest description is that the destination has become a program counter. That is why five control planes collapse into one: the path is held at the ingress edge, and the core just forwards.
one address · four instructions · rewritten at every hop
0200 getting there
Reversible, and it strands nothing.
Plan against the live RIB
The addressing plan is proposed against the routing table you actually have, and checked for reachability before anything is written. The check is a proof, not an opinion.
One write path
Every change to the network goes through the same proposal, verification and approval sequence. The gate is mechanical, not procedural.
Reversible at every step
Each step is independently useful and independently undoable. Stop after the audit and you still keep the output.
The transition case is simplification, not renumbering.
e000 End.DT6 deliver
See the plan against your own core.
The entry is still the read-only audit on what runs today. 30–60 minutes, no change window, and the arithmetic is computed on your estate rather than ours.