On Mar 20, 2017, at 8:48 PM, Viktor Dukhovni <ietf-d...@dukhovni.org> wrote: > FWIW, when adding DANE support to Postfix,
Viktor, forgive me, but this is such a completely different use case than what we are talking about. In this case, the Postfix mailer and the recursive validating resolver are both operated by the same entity. You can use TSIG and be satisfied, or even trust the IP address if you are naive. The homenet use case is completely different. Here we are talking about devices that routinely roam among operational domains with no basis for trust or even knowledge of the trustworthiness of the local resolver. In this case, trusting the local resolver could be anywhere from completely safe, to an opportunity to have your butt sniffed by your employer, to a completely corrupt Wifi hotspot that's been set up specifically to attack your device. Trusting the network in this case is simply nonsensical, and there is no trust relationship that could make TSIG work.
_______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
