Hi, so I think the scenario you describe is worth considering in the draft.
IMO it makes the need for such a draft even more compelling because the idea of a browser sending user’s browsing data to any number of (frequently changing) third-party resolvers brings up all kinds of issues around privacy, informed user consent, not to mention unexpected and unwanted behaviour. Neil Sent from my iPhone > On 11 Mar 2019, at 13:09, Stephen Farrell <[email protected]> wrote: > > > Hiya, > >> On 11/03/2019 09:25, Neil Cook wrote: >> What other resolvers would those be? Firefox only uses Cloudflare at >> the moment. You can manually change that if you know about a >> different DoH server. > When I briefly played with FF nightly and DoH, it was > using both the system resolver and CF. I had to muck > about some to get it to actually use the CF results > because of entries in the system resolver cache. > > My point though was that the dichotomy in Vittorio's > draft is too simple - I'd guess its likely that a > browser, if doing DoH for real, would end up trying > various DoH servers as well as the system resolver > and would make possibly complex choices based on the > queries, answers and metadata (e.g. speed) seen. And > those choices could of course change as the browser > s/w is updated (all that being in addition to the > kind of user-driven config stuff Vittorio's draft > mentions). > > Cheers, > S. > <0x5AB2FAF17B172BEA.asc> > _______________________________________________ > dns-privacy mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/dns-privacy _______________________________________________ dns-privacy mailing list [email protected] https://www.ietf.org/mailman/listinfo/dns-privacy
