Summary

  • Clara Wade said she joined Tobias Fiebig’s RIPE 2024-01 IPv6 PI proposal as a coauthor to edit it after community feedback. At RIPE 91 she described cutting the draft to six pages, moving the End Site definition to a separate proposal and simplifying section 7.1.
  • The proposal reached version 3.0 in August 2026, then was withdrawn in September after feedback. Its editing history shows how much policy work happens before anybody can settle the underlying operational choices.

A bowl of spaghetti is a useful slide. It is less useful as a rule for allocating IPv6 addresses. At RIPE 91, Clara Wade borrowed the image from an earlier presentation by Tobias Fiebig, original author of a proposed revision to provider-independent IPv6 assignments. She had joined the proposal as a coauthor to edit it after the community had got its forks into the text.

Wade described three cuts. The definition of an “End Site” had become a debate big enough to occupy its own proposal, so she separated it. She simplified section 7.1 on assignment size and removed section 7.1.3, whose detailed instructions risked telling the RIPE NCC exactly how to administer existing assignments. Then she took out repetitions. The draft came down to six pages. People had told the authors they could not spare the time to read the longer version; six pages gave them a fighting chance.

That is an unusually visible piece of editorial labour in address policy. A line about what counts as a site can decide whether a network fits within a provider-independent assignment. A sentence about growth can determine how much space an applicant may receive and whether an existing block must be renumbered. Remove the wrong words and the policy loses a boundary. Keep every qualification and the people expected to discuss it may never reach page seven.

The remaining RIPE 2024-01 text still made concrete choices. It proposed a /48 starting size for IPv6 PI assignments, with justified growth at neat four-bit boundaries such as /44 and /40. It recommended reserving contiguous room for later expansion. If that room was unavailable, a replacement assignment and a six-month renumbering period formed the proposed route onward. The RIPE NCC’s impact analysis described high implementation work and pointed to requirements it would struggle to verify. It also warned that partial transfers could break up the growth space the draft hoped to reserve.

Wade’s public biography on the RIPE 90 Programme Committee page makes her interest in the process less mysterious. She said she entered the community in 2018, aged 22, while working with increasingly frequent IPv4 transactions in Buenos Aires; keeping those transactions safe and aligned with address policy became a personal project. By the time she joined this IPv6 draft, she had worked on AWS address strategy and participated in RIPE community roles. She knew policy as something people have to read, use and explain to one another.

A shorter draft did not settle every dispute. RIPE’s proposal record dates withdrawal to 23 September 2026; the RIPE NCC announcement says the proposers withdrew it based on feedback. The record gives no licence to pin that decision on one paragraph or one person. Wade had made the menu legible. The table still had to decide what to order.