Moin, am 06.11.25 um 15:42 schrieb jordi.palet--- via address-policy-wg:
[…] In my experience […] when an ISP has difficulties to obtain the prefix size they believe they need for their network […] they don’t bother to keep trying to fight with the RIR, they pass "the ball" to the customers: So instead of providing them what they initially though is the right prefix size, for example /48, they will reduce it to a /56 or even worst /60. […] We tried to fix the policy a few times to make easy the justification, and didn’t worked. I will be happy if we can find a magic wording for that, but across many years was not fixable. […] 3) Allowing /28 will allow more LIRs to provide /48 to end-sites instead of longer prefixes which is really bad. […] 4) Allowing /28 will solve the problem for a few or many, I really don’t think “to how many" is the key.
If amending the policy is not having the intented effect towards more liberal allocations, maybe it's not the policy text that is in error? Then, DTAG got a /23, so obviously it's possible to get a smaller prefix based on (supposedly) documented need. But which end-site has a use case for 65k /64 networks? DSL, Cable and Fiber ISPs at least over here hand out /59 to /56 to (residential) customers, i. e. 32 to 256 possible /64 networks per (residential) CPE. After all, I don't see a compelling reason for the proposed change; if the issue is a too strict handling of requests for a smaller prefix than /29, then explicit wording that this is ok with the community if the need is documented needs be added. Therefore still opposing this policy proposal, -kai