[dnsext] perhaps we should reintroduce "resimprove"

based on the renewed interest in the delegation and glue ttl problem
caused by the "ghost domains" paper, i looked again at:

http://tools.ietf.org/html/draft-vixie-dnsext-resimprove-00

...which i presented in prague about a year ago. the sticking point was:

    B. Stopping a downward cache search when an NXDOMAIN is encountered.

and all of section 3. this proposal was considered controversial since
two existing implementation (rbldnsd and tinydns) currently send
nxdomain when queried for an empty nonterminal domain name. i did not
agree that this was a problem since RBL DNS queries are always full
length (that is, for all octets or all nybbles of an inverted host
address) and since the DNSSEC specification clarified non-terminal names
as existing but empty.

i now propose that we dust off this draft, remove (B) and section 3, and
progress it not as an improvement but as a security and resiliency
requirement (so, a proposed standard) in the face of the "ghost domain"
problem.

i may yet reintroduce the NXDOMAIN matter but i don't think that we
should logjam on it any further.

with five shows of support i would consider the editorial work involved
here to be worth doing.

paul

_______________________________________________
dnsext mailing list
dnsext@ietf.org
https://www.ietf.org/mailman/listinfo/dnsext