Hi Hannes, About the draft-fossati-tls-attestation-01, here is my comments.
Issue 1：Position of TLS Client. I think this draft could clarify if the TLS Client in Attester is inside the RoT or outside the RoT. If the TLS Client is inside the RoT, then the TLS channel is a fully trusted point-to-point channel between TEE and Verifier/Relying Party. If the TLS Client is outside the RoT, then the TLS channel is not that trsuted. Because Man-in-the-Middle attack could intercept the message transmission between RoT and TLS Client. Even though use nonce to fresh the PAT, it can only prove the TLS Client have possession to the RoT but cannot make sure the TLS client is deployed in the RoT-related device.
Issue 2：Better add one customized attestation type for expansibility of this protocol like opaque strings. It is possible that users would like to use some customized format to express evidence or attestation result.
BR Penglin _______________________________________________ TLS mailing list TLS@ietf.org https://www.ietf.org/mailman/listinfo/tls
