/
19-July-2025 EDHOC-PSK interop testing
Edit

19-July-2025 EDHOC-PSK interop testing

Page to annotate the results of the interoperability test of EDHOC-PSK during the IETF-123 Hackathon.

📃NOTES

  1. According to RFC9668 (EDHOC over CoAP), message_1 needs to be preceded by a True value 0xF5. https://www.rfc-editor.org/rfc/rfc9668.html

  2. In message_3, c_r is prepended so that the server can verify that the message is part of the handshake. (Part of CoAP profile, RFC9668)

  3. Label KEYSTREAM_3A was incorrect (it changes in EDHOC-PSK with respect to EDHOC RFC9528)

  4. TH_4: Do we use CRED_PSK or key_CRED_PSK?

    The issue here is that the CRED_PSK, if defined as the whole structure, contains as well the sub field, which will be different for I and R in order to avoid selfie attacks. Therefore, if we consider CRED_PSK as the whole structure, values such as PRK_4e3m and TH_4 will differ. It might be, therefore, useful to define a new label key_CRED_PSK to refer to the /k/ field (key) of the COSE_key field of the CRED_PSK

    CRED_PSK {                                               /CCS/   2 : "mydotbot",                               /sub/   8 : {                                         /cnf/     1 : {                                       /COSE_Key/        1 : 4,                                   /kty/        2 : h'0f',                               /kid/       -1 : h'50930FF462A77A3540CF546325DEA214'  /k/     }   } } key_CRED_PSK: 50930FF462A77A3540CF546325DEA214

  5. How to define ID_CRED_PSK in message_3? Since we are using the optimization for encoding, we have to follow https://www.rfc-editor.org/rfc/rfc9528.html#name-compact-encoding-of-id_cred . In our example, the value is 0x10, so we keep it as it is, no need for adding 0x41 etc.

✅SUMMARY

The main points are:

  • Update to the latest version of the draft and follow RFC9668 that specifies how to use EDHOC over CoAP.

    • Mistakes included using incorrect labels, incorrect versions of the draft (especially in k_3/iv_3 or prk_4e3m)

  • General comment: We need to clearly define what is CRED_PSK, and whether we use the whole credential or just the key of the COSE_key field. (This affects TH_4, PRK_4e3m)

💻IMPLEMENTATIONS