On Tuesday, 26 May 2020 23:56:30 UTC Ben Schwartz wrote: > On Tue, May 26, 2020 at 7:06 PM Paul Vixie <p...@redbarn.org> wrote: > ... > > > and the responder knows > > the additional data (that is, it should not make extra queries to find it > > just > > for additional data reasons.) > > The current text does not have this caveat: it recommends that the > recursive perform those extra queries before replying, because they will > otherwise have to be performed by the client at higher cost.
i think that's overly prescriptive. the client may already have the data in private cache from its prior work, or it may be striping its DNS queries across several full resolvers and therefore able to get the data faster if you don't provide it than if you take the time to fetch it before you answer. > However, it > also says "recursive resolvers MAY ... produce a reply that omits some of > the associated RRSets ... when responding before fully chasing dependencies > would improve performance", so a compliant recursive resolver could > certainly implement the behavior you're describing. nice. that's all the rope i'd ask for. -- Paul _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
