Summary

  • A trial approved in 2001 used historically allocated IPv4 space to develop IPv6 services and was scheduled to finish at the end of 2005. Kosuke Ito and three coauthors proposed extending it through 2008 because working services still relied on dual stack.
  • APNIC 19 recorded consensus after questions about a contractual return, the full /8 block and whether generous IPv4 access weakened the incentive to build IPv6 applications. A 2006 report shows one participant returned space to the IPv6 Promotion Council; the others continued in phase two.

The exit date arrived before the exit technology. The IPv6 Promotion Council of Japan had spent four years letting participating networks use a large block of historically allocated IPv4 space while they developed services for an IPv6 future. In the January 2005 proposal, Kosuke Ito and his three coauthors admitted that the original end-of-year handback depended on an optimistic forecast. IPv6 had not displaced IPv4 quickly enough. Actual services ran both stacks.

Their request was precise: move the trial's finish from the end of 2005 to the end of 2008. Each participant would need an IPv6 deployment plan, a renewed agreement and periodic reports. The proposal's list of disadvantages read “none.” That was a neat line for a complicated bargain: the trial intended to promote the successor protocol by continuing to use the scarce predecessor.

The APNIC 19 Policy SIG minutes show what the room made of it. Ito and Yoshiyuki Ezura presented the extension on 24 February. Someone asked for a firm contractual duty to return the IPv4 addresses; the presenters confirmed it. Another intervention demanded that the entire /8 return to the public pool, rather than leave participants with a privileged route around ordinary address policy. A third objection went to the incentive itself: if developers could keep building on ample IPv4, why hurry to build IPv6 services? The cochair called for hands twice and recorded consensus.

Consensus changed the shape of the permission. The meeting's summary paired the 2008 end date with IPv6 service plans, return of the IPv4 space, regular public reporting and agreements that specified those conditions. APNIC's proposal history records Executive Council endorsement on 13 May 2005. It records approval, not the eventual disposition of every address in the block.

Ito's March 2006 activity update supplies a narrower result. Phase one had closed at the end of 2005. One participant that had moved to IPv6 returned its address allocation to the IPv6 Promotion Council by the deadline. Other participants carried their allocations into phase two, which began on 1 January 2006. They were to set IPv6 goals and schedules and report twice a year. The update also named equipment, access-network and software barriers. The extension was not an abstract plea for patience; it was a response to missing operational pieces.

The 2005 agreement put a date on a promise. The 2006 report shows one partial return and a continuing trial. A record of the full /8's return to APNIC at the end of 2008 would settle the larger handback question; these sources do not contain it. The immediate lesson is already visible: a sunset date on a registry's calendar cannot turn dual-stack customers into IPv6-only customers by decree.