I have three main concerns on the draft:

1. Precisely which keys are compromised and how? I believe the "how"
   part is necessary to check that the solution actually solves the
   problem. (e.g., application_traffic_secret being compromised because
   of compromise of g^xy is different from application_traffic_secret
   compromised without the compromise of g^xy)
2. Having security considerations with only "This entire document is
   about security." is unacceptable.
3. Terminology in the draft is quite imprecise. Many terms have been
   invented, such as "SEND keys" but never defined clearly. From
   terminology perspective, please stay as close to 8446bis as
   possible, and properly define any new terminology in the terminology
   section. I also don't like "initiator" because by default I always
   think of that as Client in TLS context, but the draft uses it for
   server also.

More detailed comments and some concrete suggestions are in the issue [0].

-Usama

[0] https://github.com/tlswg/tls-key-update/issues/59

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

_______________________________________________
TLS mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to