Op 27-05-2026 om 01:39 schreef Brian E Carpenter:

Which is true, as far as I can tell, of *any* double encryption, as long as the 
two algorithms are strictly independent.


Hybrid key exchange is not the same as double encryption. Hybrid key exchanges 
also require both *keys* to be independent, and the combiner function to be 
theoretically sound, and the combiner function to be implemented correctly. All 
of which I have seen go wrong in real-world code. So the probability of the 
hybrid being broken is really pq + r, with r the probability of any of those 
accidents happening.

And hybrid *signatures* (which is what this last call is about) are even 
trickier. As an example, the hybrids from draft-ietf-lamps-pq-composite-sigs 
are only EUF-CMA, even though ML-DSA by itself is SUF-CMA. Now every protocol 
designer using those constructions needs to figure out whether their protocol 
needs SUF-CMA for security or is still ok with EUF-CMA. Good luck with that. So 
there are good reasons to prefer pure ML-DSA over hybrid signatures.

----

Marc Penninga

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

Reply via email to