You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The recipient MUST NOT tell the reason for decryption error in suit-report
even if the recipient encrypts it, some attacker may guess the reason from ciphertext of suit-report
Do we have to mention it?
Yes. It is better to add some text to suit-firmware-encryption or suit-report document like "The recipient MUST NOT tell the detailed error message such as padding incorrect to the sender in suit-report."
Decryption Oracle Attacks
What is it?
The attacker can obtain the plaintext without knowing the encryption key
if the attacker knows the plaintext pattern
It could be practical if there is a low entropy template for plaintext payload
The sender MUST include the algorithm id into the value of suit-encryption-info for AEAD algorithms
The recipient MUST verify it before decryption e.g. by verifying the Manifest block in Envelope with suit-authentication-wrapper
Countermeasures against Decryption Oracle Attacks with AES-CBC
Separate AES keys of -CBC and -GCM if the recipient and sender support both of them (from the AEAD-to-CBC Downgrade Attacks on CMS slide)
(I think we need to research more and more.)
Do we have to mention it?
Against Downgrade Attacks: It might be better to add "The algorithm identifier MUST be placed in protected or unprotected header." somewhere in the suit-firmware-encryption.
In an oracle attack, an attacker queries a system with a cryptographic task and observes a function of the task’s outcome. If the function validates the decrypted plaintext’s padding, it is called a padding oracle [3]. More generally, if the function checks the plaintext’s format, it is called a format oracle [29].
The text was updated successfully, but these errors were encountered:
kentakayama
changed the title
No mention against Decryption Oracle Attacks a
No mention against Decryption Oracle Attacks
Jan 30, 2024
kentakayama
changed the title
No mention against Decryption Oracle Attacks
No mention against some Oracle Attacks
Jan 30, 2024
@hannestschofenig
In the last meeting, we've talked a bit about the Decryption Oracle Attacks presented at LAMPS WG in IETF 118.
While researching on it, I found another attack, Padding Oracle Attack, on AES-CBC mode and it seems more realistic.
I think it is better to add some notes in security considerations in suit-firmware-encryption or suit-report documents.
I'm still wondering the Decryption Oracle Attacks is realistic on SUIT Encrypted Payload case.
The recipient SHOULD carefully reply on decryption failure to prevent some attacks.
Padding Oracle Attacks (+ Encryption Oracle Attacks)
padding incorrect
to the sender in suit-report."Decryption Oracle Attacks
suit-encryption-info
for AEAD algorithmssuit-authentication-wrapper
Resources
The text was updated successfully, but these errors were encountered: