> bert hubert <mailto:bert.hub...@netherlabs.nl> > Tuesday, March 17, 2015 12:05 AM > > Sorry? We solve implementation hardship by standards action now?
as with client-subnet, we recognize that people will do what they want, or stop doing what they don't want, especially if they are CDN providers with a lot of revenue and a lot of expense riding on their choices. i don't love this situation but i can tell you that quoting specifications at folks and using words like "mandatory" isn't the way to change their minds (or their deeds.) noting that there's a more-than-ten-years-old CNAME patch to qmail that just about everybody is supposedly running, i expect the operational impact of phasing out ANY to be ~0. also, a lot of operators foolishly patched their BIND servers to stop answering ANY because they considered it a DDoS risk (which is patently insane but please don't shoot the messenger) and not a single qmail user was heard from on the matter. the internet works by cooperation. often, first mover advantage is sticky. but almost as often, somebody like the mozilla dev team decides that something like ANY is the solution to their API layering problem, and the rest of us rip the bandaids off and study the underlying wound. so it is in this case. now, mozilla has backed off, but the underlying wound remains a visible topic of conversation. to me the use case is, it's an information leak, and i don't want my cache probed, and i can't tell the difference between a cache prober and qmail, so into the same stew pot they both must go. (along with RD=0 on an RA=1 server.) -- Paul Vixie
_______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
