You're right, this is all there in the draft. It's just scattered around a bit, and "anonymity set" is used only once and not defined.
I filed an issue https://github.com/tlswg/draft-ietf-tls-esni/issues/204 in case the editors want to consolidate text on this concern. thanks, Rob On Mon, Nov 25, 2019 at 11:25 AM Ben Schwartz <bem...@google.com> wrote: > The record_digest, like the ESNIConfig itself, is intended to be constant > across all domains that form an anonymity set (i.e. O(1) ESNIConfigs per > CDN). Thus, the record_digest reveals no additional information to an > onlooker who can observe the server IP. > > On Mon, Nov 25, 2019 at 2:03 PM Rob Sayre <say...@gmail.com> wrote: > >> Hi, >> >> I see the issue of tracking and blocking via record_digest has come up a >> few times in the github repository, but I don't understand the resolution. >> >> If someone wanted to observe or block traffic to "example.com", couldn't >> they retrieve the ESNI keys, calculate the record_digest themselves, and >> then use that to spot traffic to "example.com"? >> >> Is the idea that DNS providers will vary the shared keys? >> >> thanks, >> Rob >> >> _______________________________________________ >> TLS mailing list >> TLS@ietf.org >> https://www.ietf.org/mailman/listinfo/tls >> >
_______________________________________________ TLS mailing list TLS@ietf.org https://www.ietf.org/mailman/listinfo/tls
