> --On Wednesday, 11 October, 2006 21:59 +0200 Stephane Bortzmeyer > <bortzmeyer@xxxxxx> wrote: > > On Wed, Oct 11, 2006 at 01:03:24PM -0400, > > Keith Moore <moore@xxxxxxxxxx> wrote > > a message of 28 lines which said: > > > >> In the past month or so I've run across two separate ISPs > >> that are apparently polluting the DNS by returning A records > >> in cases where the authoritative server would either return > >> NXDOMAIN or no answers. > > > > Today, it is quite common and it becomes more and more common. > > > >> Is there anything that IETF as an organization, or IETF > >> participants, can do to discourage this? > > > > Producing a RFC 4084bis is, IMHO, the best way to go. > > Currently, RFC 4084 does not address this issue, only a > > related issue: > > > >> o DNS support. > >> Are users required to utilize DNS servers provided by > >> the service provider, or are DNS queries permitted to > >> reach arbitrary servers? > > > > So, there is IMHO a good reason to upgrade the RFC. > I think there are several other reasons as well. If we were > doing 4084 over again based on what I think we know today, I'd > recommend putting less emphasis on email issues --or even moving > them to a separate, supplemental, document-- and doing more work > on DNS tricks, the behavior of hidden and/or mandatory proxies > (if I recall, 4084 doesn't go further than a requirement that > they be identified), and preferential treatment of customers, > content providers, or applications (more or less the "net > neutrality" discussion, plus some issues about who gets hurt if > QoS options are provided to some customers and the relevant > network starts getting short of capacity or other resources). > But, while I could get up the energy to cheer if someone else > felt like doing the work, I'm personally disinclined to reopen > 4084, or to try to persuade the IESG to do so, unless someone > can show where it is actually being used enough to do some good > and/or provide a persuasive argument about where it would be > used if identifiable changes were made. Agreed on all points. But another and IMO more serious issue with attempting to address this in RFC 4084bis is that a document entitled "Terminology for Describing Internet Connectivity" isn't where I'd be inclined to look for rules regarding DNS operations. IMO this is a sufficiently serious issue that it needs to be dealt with in a separate document that addresses it specifically. I'd be happy to try and contribute to such a document. Ned _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf