John Mattsson <[email protected]> writes:

> With well-designed composite KEMs, you can indeed obtain the best of
> both worlds, i.e., security corresponding to Max(PQ, T). However, the
> "T" component of PQ/T should soon no longer be considered to provide
> meaningful security, particularly for key exchange. In that case, the
> effective security becomes Max(PQ, 0), which is simply PQ.

The above argument is often repeated, but I think there are naunces that
get lost when phrased like that.  Security is rarely binary either or,
but more of a spectrum.  All ECDSA keys in the world won't automatically
be revealed on the first day a CRQC is demonstrated.  People still run
RSA 1024 deployments (e.g., DNSSEC) , and will continue to do so even
after 2029.  In their threat model the crypto risk is not the biggest
problem, and RSA-1024 provides >0 security and is not the weakest link.

We won't have sufficient confidence in practical real-world uses of
ML-KEM-512 or ML-DSA-44 until say 5-10 years after initial deployments,
which is only starting now.  Throwing away all >0 traditional security
which non-hybrids would offer is a HNDL/MITM risk.

/Simon

Attachment: signature.asc
Description: PGP signature

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

Reply via email to