Summary

  • The 26 October revision of an IETF IANABIS working-group draft adds a rule of thumb: an early IANA allocation should stay out of a published RFC while it remains temporary.
  • The draft would give early allocations a two-year initial term, but an expired value can later be deprecated or deallocated. An RFC paragraph has no matching expiry switch.

A number looks authoritative once it appears in an RFC. In the new early-allocation draft, the IANABIS working group has put that visual certainty on a short leash. Its newly added Section 3 sentence tells authors to keep early allocations out of published RFCs: an allocation allowed to expire may be deleted or deprecated. The July revision had no such sentence. This is a fresh instruction about where a provisional number belongs, rather than a change to the number itself.

Early allocation solves a real coordination problem. Independent implementers need the same protocol value while a specification is still a draft. IANA can reserve one and mark it temporary. The reservation stops competing experiments from picking incompatible numbers. It cannot make the draft finished. Under the current RFC 7120 process, that reservation has a clock; renewal and later disposition are separate decisions.

Revision 04 proposes a two-year initial term in place of the current one-year term. IANA would publish the expiry date while the draft moves toward IESG review. If the draft reaches the IESG with a valid allocation, the draft says the allocation stays live during that consideration. Otherwise the clock matters: a value may be renewed, deprecated, deallocated or left in its recorded state after the applicable checks. The precise route varies with the registry and request. A printed RFC cannot be silently edited to follow each turn of that registry record.

The newly added sentence therefore assigns the risk to the right stage of publication. A temporary registry row can help testing today. A permanent standards document should wait for the final assignment before treating that row as settled text. The draft itself still needs approval before this guidance becomes a Best Current Practice. Its next revision will show whether the warning survives and whether it specifies how authors should handle an early value already cited by another publication.

No named RFC and IANA history in this source package shows an actual citation that later went stale. Such a pair would put a measured repair bill on the problem. The proposal already identifies the failure path: a provisional number migrates from a changeable registry into a document built to stay put.