Summary

  • Tomoya Yoshida proposed joint RIR and ISP testing so newly allocated address blocks could be checked for reachability, not merely entered in a registry.
  • APNIC 22 backed IPv6, automatic tests and information sharing but found no consensus for an IPv4 rule; the revised proposal was abandoned after final call.

The registry can hand an operator a fresh address block. It cannot make every router on the Internet forget an old bogon filter. Tomoya Yoshida's 2006 prop-039 took aim at that awkward interval between allocation and use. He pointed to trial tests of 121/8, 122/8 and 123/8, and wanted a full service for testing new IANA blocks, notifying ISPs and coordinating across regional registries. The paper asserted that many new blocks remained unreachable, some for as long as a year. It supplied no list or denominator for that assertion. A measured series across named vantage points would show how widespread the problem was.

At APNIC 22, the room did something more revealing than a simple yes or no. The chair found support for the de-bogonizing project in principle, an IPv6 rule, automatic ICMP or traceroute checks with notification, and sharing results among RIRs. There was no consensus for an IPv4 rule. Participants questioned whether benefits would justify costs and whether tests could reach the real source of stale filters near network edges. APNIC's own difficulty getting replies from filtering ISPs was part of the discussion. A registry could detect a dark route; it still needed an operator to change a filter.

Yoshida's second version kept the elements the meeting had accepted and dropped those it had not. It proposed a defined cross-RIR service, ISP test servers and common IPv6 prefix guidance, with implementation details left to the registries. This was a narrower offer, not a deployed service. APNIC's proposal history records an eight-week final call extended by two weeks after unresolved discussion. Consensus still failed, and prop-039 was abandoned on 5 December 2006.

The half-consensus matters. People agreed that a registry record was too weak a test of an address block's usefulness. They did not agree on the institution, scope and continuing work required to turn that observation into policy. Yoshida's proposal made the gap visible; the final call left it without his proposed rule.

Portrait identity reference: APNIC 26's named speaker listing for Tomoya Yoshida.