IPv6 mistakes, was: Re: Looking for an IPv6 naysayer...

Iljitsch van Beijnum iljitsch at muada.com
Fri Feb 18 11:16:09 UTC 2011


On 18 feb 2011, at 12:00, Patrick W. Gilmore wrote:

>> How can they "return" stuff to ARIN that they got from IANA in the first place?

>> ARIN seems to be getting the very long end of the legacy stick.

> But last time I checked, the United States is in the ARIN region.  And ARIN did not exist when the US DoD got its space.  (In fact, I do believe the reason "IP space" exists is because the DoD paid someone to come up with the idea? :)

True, but how is all of that relevant?

> If the US DoD wants more space, it has to ask ARIN, right?  Are you suggesting it should deal with a different organization depending on which direction the IP addresses flow?

> Supposed it was space ARIN assigned the DoD?

Policies like giving each RIR one of the final five /8s were carefully created to give each RIR equal access to address space. Automatically giving legacy space to the RIR for the region that the holder of the legacy space is in is incompatible with that, and means that ARIN will get virtually all of it.

To me, it seems both natural and fair that legacy space (especially /8s) is returned to IANA and then redistributed over the RIRs.

By the way, IANA only deals in /8s. However, a lot of people got legacy /16s or other non-/8 sizes, so some /8s that are marked "legacy" actually contain a lot of unused space. Each of those /8 is "administered" by a RIR, but it's unclear (to me at least) whether that means that RIR gets to give out that space in its region or not. And if not, what is supposed to happen with this space. It's a significant amount, about half the size of the class E space:

RIR      Administerd by        Delegated           Free

afrinic         33.55 M           8.71 M        24.85 M
apnic          100.66 M          77.95 M        22.72 M
arin           671.09 M         592.04 M        79.05 M
ripencc         67.11 M          63.01 M         4.10 M





More information about the NANOG mailing list