At 13:22 20/07/2007, Peter Koch wrote:
Dear WG,
the draft <draft-ietf-dnsop-respsize-07.txt>, "DNS Referral Response Size
Issues" has been on our plate for quite a while. After the Prague meeting,
four people have come forward with a review, all but one supported the
document with minor changes suggested, the fourth reviewer put more strength
in his concerns which were focused on particular paragraphs, not the
overall structure or direction of the draft. There was not much discussion
or contention w.r.t. the proposed text changes.
We have still three people left on our WGLC volunteers list, so to save
another editing cycle, this message starts a
Working Group Last Call on <draft-ietf-dnsop-respsize-07.txt>
ending Wednesday, 2007-08-22 12:00 UTC
Please review the draft and the list thread starting 2007-05-07 and comment
on this list, preferrably including proposed text if you'd like to see
changes. Our charter lists this document as aiming at "Informational"
status.
We will have one more version of the draft to incorporate current
and WGLC text
changes and NITS review, the -08 to go to our AD.
Thanks,
Peter
Review:
General:
Mostly OK, but some rough edges, it is not clear if this is informational
only document or if it is making some recommendations.
The document does not cite 2119, and does not use MUST/SHOULD in upper case
but some of the statements read like upper case is indented.
Section 1.2 (issue)
I think this section is out of date, most recursive resolvers support ENDS
by now. In a quick sample I did on my authoritative nameserver logs
I found almost 20000 different addresses that asked my server questions
during 12.25 days earlier this month.
over 86% of outside queries have ENDS0
over 79% of outside addresses sent at least one ENDS0 query
only 1% of outside addresses are inconsistent in ENDS0 use.
If the section is trying to address lack of support by stub-resolvers it
should say that explicitly.
Section 1: (Nit)
I think section 1 should point out that ENDS0 support has been recommended
for a long time, at least since RFC3226 (dec 2001)
Section 2.1.2 (nit)
s/512 octets/512 octets or advertised space in ENDS/
Section 2.2 (question)
It does not explicitly say that zone owners SHOULD/MUST run ENDS0
capable nameservers or recursive resolvers.
Is this intentional ?
Section 2.3.2 (nit)
I think this one belongs in section 2.2 as this is more appropriately
guidance to zone owner.
Section 5. (nit)
I think this should be appendix A.
Olafur
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www1.ietf.org/mailman/listinfo/dnsop
