Hi Wilfried, Sorry for not responding to this sooner, job transition and time off work didn't help... Wilfried Woeber, UniVie/ACOnet said the following on 9/08/11 19:49 :
I agree, and thus I support the proposed policy.
Thank you! :-)
Checking the text again, I do see some aspects that may be ambiguous and/or require clarification:
Section 1.0, after bullet point 2, <quote> The Recovered IPv4 Pool will stay inactive until the first RIR has less than a total of a /9 in its inventory of IPv4 address space. </quote>
As the different regions do have, or are still discussing, various (micro-)policies to reserve blocks or to set aside blocks for specific purposes, the term "inventory" either needs clarification/definition or there should be an explanation which part(s) of the RIR's IPv4 address space is considered to belong to the "inventory".
My personal proposal would be to clearly state that *all* currently unused space in an RIR is considered its "inventory".
The authors intention was that all currently *unused* space would be considered to be the RIR's inventory.
As an aside, and given the fact that there is no specification about the fate of returned address blocks within a Region, what is expected to happen, when and if no RIR is below the /9 threshold any longer? Will the RIP (Recovered IPv4 Pool :-) ) be declared inactive again, or not? I do understand that this is not very likely.
The intention of the authors was that this policy remains until replaced by another policy. Otherwise we'd have written that this policy would no longer apply once no RIR was below the /9 threshold again.
Regarding timelines:
It is not clear to me, whether the 1st round of redistribution happens "immediately" after declaring the RIP active, and thereafter will be synched to the March1/Sept1 schedule? Alternatively, the 1st round may be executed at the next scheduled date.
My personal preference here would be to act "immediately" and then take up the schedule, with the effect of potentially not having anything to dish out for the next round(s) and thus skip the next scheduled date(s) .
However, this approach may run against the intention(?) of proposing a fixed 6m schedule (in order to accumulate a reasonably big RIP?).
Maybe the authors could explain the rationale for the particular schedule?
We picked two months of the year, and March/Sept seemed reasonable in that they avoided calendar anniversaries and some major holiday seasons. Hope this helps! philip --