On Fri, Apr 3, 2020, at 12:50 PM, Stephen Farrell wrote: > > Hi Joe, > > Pre-agenda bashing question for chairs and authors of > the ESNI/ECHO draft... > > I'd like to try again to suggest simplifications for > ECHOConfig (e.g. no extensions, just one public key > per RR VALUE etc.) and perhaps also some more rules > constraining inner/outer CH variances. > > How would you prefer those be raised? As issues in GH > or threads on the list and would you rather I try group > 'em as much as possible or make each issue as fine-grained > as possible?
Fine-grained issues on GitHub would be my preference! Thanks in advance for filing them. > Optimistically, this meeting might be the one where we > mostly tie this stuff down, so I hope it's ok that some > of these issues have been raised before in one form or > another. The move from ESNI->ECHO and to HTTPSSVC and > the goal to get to WGLC I think justifies re-checking > if we really need the current level of (what I claim is > too much:-) complexity in ECHO. Hmm... the issues you note above seem somewhat orthogonal to these two changes. The ESNI->ECHO change doesn't seem to affect ECHOConfig contents or extensibility. (We effectively renamed the struct while making the change.) Similarly, HTTPSSVC has more or less the same extensibility properties as the old ESNI RR. That said, I think if there's new data or information that suggests we collectively made poor design decisions, raising them would be helpful. I'll leave it to Sean and Joe to determine whether or not re-hashing some of these issues is appropriate. Rules around inner/outer CH variance are ripe for discussion, though! Best, Chris (no hat) _______________________________________________ TLS mailing list TLS@ietf.org https://www.ietf.org/mailman/listinfo/tls
