Summary

  • James Spenceley and Jonny Martin's prop-070-v001, dated 23 January 2009, proposed a maximum size for a single APNIC IPv4 allocation: initially /8, then /9, /12, /14, /16 and finally /18 as the IANA free pool crossed specified thresholds.
  • At APNIC 27, supporters argued against one large applicant taking the last available space while smaller networks waited. Participants also raised customer demand, regional disadvantage, repeat-application work and NAT incentives. APNIC records the proposal as abandoned on 26 February 2009.

The last large customer in line would not see an equation. It would see APNIC say that its justified request was bigger than the maximum permitted that week. Spenceley and Martin wanted that ceiling to fall as IPv4 ran out. Their January 2009 proposal made global scarcity an automatic limit on an individual APNIC allocation, even when the applicant could demonstrate need.

The original ladder begins with a /8 cap while IANA holds 30 blocks of that size. At 25 remaining, the cap becomes /9; at 20, /12; at 15, /14; at 10, /16; and at five, /18. In address counts, those endpoints are 16,777,216 and 16,384—a thousandfold reduction in one request's ceiling. The thresholds concern IANA's global free pool, not a tally of addresses inside a particular applicant's network. The separate final-/8 reserve was outside this proposal's scope.

Spenceley's worry was a distribution cliff. In the APNIC 27 transcript, he described a large member or national registry claiming the remaining space ahead of 20 or 30 smaller members. The written proposal does not identify an applicant who had done that. It offered an insurance policy against the last big request, not a remedy for a measured incident. The authors themselves named two costs: an APNIC-only ceiling could disadvantage the region while other registries issued larger blocks, and smaller allocations could force members and staff through more applications.

The room made the hidden cost visible. Izumi Okutani asked what APNIC should tell a large operator whose customers required more addresses than the cap allowed. John Schnizlein worried prolonged rationing could push providers toward more NAT and walled gardens. Martin said the hard maximum looked unlikely to pass and floated a queueing alternative. Philip Smith pointed to repeated application work. Spenceley agreed that the design had become complicated, but still valued a guard against one member taking a surprisingly large share of the pool.

He did not ask the meeting to declare consensus for the hard cap. The official proposal history marks prop-070 abandoned at APNIC 27 on 26 February. No member was actually held to the /18 ceiling by this proposal. Its failure exposes a real allocation choice: spread the last free addresses across more applicants, or let demonstrated operational demand determine the size of each award. A ladder can make scarcity look orderly. It cannot make the denied customer disappear.

Sources

Original prop-070-v001 · APNIC 27 Policy SIG transcript · APNIC proposal history · Official named photograph used for the portrait