[ppml] Is this Lame or what?
briand at ca.afilias.info
Wed Sep 12 10:51:12 EDT 2007
John Von Essen wrote:
> Hostmaster confirmed what you state below:
> "Only if all of zones for a given name server of a specific network
> registration are lame, is the delegation registration deemed lame"
> That is valid, correct, and should stay as-is.
Actually, that is, IMHO, incorrect. It is backwards.
It *should* read:
"Only if all name servers for a zone of a specific network registration
are lame, is the delegation deemed lame."
If 4 /24's are delegated, and all 4 have the same NS sets on their
in-addr.arpa. delegations, but none of them
in fact serve one of the /24 in-addr.arpa. zones, then the zone
delegataion for *that* /24 should be deemed lame.
This will fix the /24 problem.
The /16 problem is another kettle of fish.
How do we address this via Public Policy process? Since it directly
affects the NRPM, should not the
definition itself be part of it, e.g. in an Appendix?
> On Sep 11, 2007, at 7:48 PM, Scott Leibrand wrote:
>> After doing some research (which I highly recommend everyone do before
>> posting), I'm not sure any policy changes are needed to deal with the
>> two specific cases raised so far.
>> ARIN's Lame Delegations page at
>> http://www.arin.net/reference/lame_delegations.html states that "ARIN
>> tests a reverse zone by requesting the SOA (Start of Authority) record
>> from the name servers registered in WHOIS. ... Any other answer (or an
>> inability to reach the nameserver due to a forward lookup failure)
>> results in the delegation being deemed lame. If all of zones for a
>> name server of a specific network registration are lame, the
>> registration is deemed lame."
More information about the ARIN-PPML