RIPE NCC · RIR Watchdog
Start here for the proposal-by-proposal record: what changed, why it mattered, who argued for what, how positions evolved, what was decided, and where the evidence still has gaps.
Source facts
Who actually shaped the debate?
BTW Analysis
The proposal would have replaced the default /24 assignment for new IXPs with a needs-based model starting at /27 and allowing assignments up to /22. An IXP seeking a larger assignment would have had to return its existing assignment and demonstrate need for at least 50% of the new assignment within two years. The only supplied version, version 1.0, was introduced on 10 October 2019 and the proposal was withdrawn on 9 January 2020. No effective change to the policy is shown.
The central issue was whether conserving the IXP address pool justified moving away from a uniform /24 default, given the operational risk that prefixes longer than /24 may be filtered in BGP. Supporters emphasized more efficient use of scarce address space and a needs-based path to larger assignments. Opponents emphasized reachability for networks connecting to IXPs and the practical consequences of deployed filtering policies.
The recorded outcome was withdrawal, not adoption. The supplied record contains one formal author, Remco van Mook of Asteroid International, but no captured discussion messages, discussion participants, or active discussion days. Accordingly, the facts establish the proposal's lifecycle and its stated arguments, but do not establish why it was withdrawn or show a broader deliberative process.
The participation record is too limited to support a meaningful concentration analysis. It records one formal actor and zero discussion participants and messages; the reported HHI and message-share measures therefore do not demonstrate either concentrated debate or broad participation. The available evidence shows authorship, not community support or opposition.
The proposal's main operational chokepoint was the interaction between assignment size and BGP filtering. A /27 could conserve address space, but if connecting networks filter prefixes longer than /24, an IXP's peering LAN announcement could become unreachable or degraded. The policy therefore depended on operational behavior outside the assignment rule itself. A second chokepoint was the proposed needs assessment: larger assignments would require proof of at least 50% need within two years, increasing the importance of how that need would be evaluated. The supplied facts do not show how either issue would have been handled in practice.
Loading
Partial
Earliest captured: Jun 30, 1992
Latest captured: Nov 12, 2026
