Tony Finch wrote:
Mark Andrews<ma...@isc.org> wrote:
In message<caje_bqckdfhnjdvxw_a6n+w27calo8jeqxwjjfx29rbh1r_...@mail.gmail.com>,
=?UTF-8?B?56We5piO6YGU5ZOJ?= writes:
In my understanding the latest major concern is about the first
paragraph of Section 2:
When an iterative caching DNS resolver receives a response NXDOMAIN,
it SHOULD store it in its cache and all names and RRsets at or below
that node SHOULD then be considered to be unreachable. Subsequent
queries for such names SHOULD elicit an NXDOMAIN response.
It is a SHOULD not a MUST. Having a existing cache entry is a reasonable
exception to the SHOULD.
The second SHOULD in the quote above seems to be explicitly saying that
those cache entries should be made unreachable, so I think it would be
unreasonable to interpret the SHOULD as meaning "don't implement this
requirement".
(Really, I think if a spec uses words like SHOULD, it ought to include a
discussion of when it is reasonable not to implement that requirement and
what are the consequences of not doing so.)
I think the spec needs to make an explicit distinction between conforming
behaviour for names already in the cache and names not in the cache.
because of how these documents are read and interpreted, i think
requirements is the wrong word.
we should describe the positive benefits to the dns system (without
mentioning any benefit or cost to any implementor or implementation style).
"As implied by STD 13 and as made explicit herein, an authoritative
response of code 3 (NXDOMAIN) asserts the nonexistence of both its QNAME
and all possible subdomains. DNS coherence can be systemically improved
if, after receiving such a response and before receiving some new
response indicating the existence of some RRset at or below that QNAME,
a recursive server answers with code 3 (NXDOMAIN) for all questions
which are equal to or subdomains of that QNAME, even if the recursive
server has unexpired RRsets stored at subdomains of that QNAME. As to
whether the resources held by such cached RRsets are reclaimed, or
whether they are allowed to expire naturally, this specification makes
no recommendation."
very few people are going to listen to a SHOULD or a MUST here, and
those who listen will likely treat those words as equivilent. so let's
not speak either word. we should describe the positive benefits of
following this implication of STD 13, and then get off the stage.
--
P Vixie
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
