Closed davidben closed 3 years ago
I went with "known" instead of "active" since {{client-facing-server}} already used "known". It just never defined it, so I guess now we have a definition or sorts. :-)
On 04/08/2021 22:29, David Benjamin wrote:
@sftcd, how does this look to you?
That looks good, better than I'd suggested.
I guess I'll close the other one in a few minutes (just to give someone time to correct me that that's not the right button-pushing:-)
Thanks, S.
The discussion on how to allocate config_ids made it very unclear whether we cared about collisions in DNS or across client-facing servers. (The former is insufficient and the latter is practically impossible.) Instead, pull this into a separate section, so we can be clear this is a recommendation for the client-facing server and the backend server.
Also downgrade from MUST to SHOULD because collisions work fine, just don't perform as well.
It was also weird that we talked about config_id in not the section about config_ids, but in the little side note about HpkeKeyConfig. You can view, comment on, or merge this pull request online at:
https://github.com/tlswg/draft-ietf-tls-esni/pull/491
-- Commit Summary --
- Expand on config_ids.
-- File Changes --
M draft-ietf-tls-esni.md (32)
-- Patch Links --
https://github.com/tlswg/draft-ietf-tls-esni/pull/491.patch https://github.com/tlswg/draft-ietf-tls-esni/pull/491.diff
@sftcd, how does this look to you?
The discussion on how to allocate config_ids made it very unclear whether we cared about collisions in DNS or across client-facing servers. (The former is insufficient and the latter is practically impossible.) Instead, pull this into a separate section, so we can be clear this is a recommendation for the client-facing server and the backend server.
Also downgrade from MUST to SHOULD because collisions work fine, just don't perform as well.
It was also weird that we talked about config_id in not the section about config_ids, but in the little side note about HpkeKeyConfig.