I think we should all agree that this choice, which has dominated much of the 
discussion for years, is one of the least impactful aspects of the transition.

The pros and cons between the two approaches are hard to decide precisely 
because the actual security difference is pretty small, and therefore the 
result you arrive at will depend strongly on your use case and assumptions.

The consequences of fighting over the small differences, however, are real and 
substantial.

-Tim
________________________________
From: Blumenthal, Uri - 0553 - MITLL <[email protected]>
Sent: Monday, April 6, 2026 3:54 PM
To: Muhammad Usama Sardar <[email protected]>
Cc: [email protected] <[email protected]>
Subject: [TLS] Re: [EXT] Re: New Liaison Statement, "Liaison communication to 
IETF regarding draft-ietf-tls-mlkem"

It deeply surprises me that IEEE is starting off its PQC transition with
non-hybrids rather than hybrids. If they have done no analysis, we
should tell them the risks and that hybrids are currently preferable.
That's why I keep emphasizing that we should first recommend hybrids and
that risks should be thoroughly mentioned in pure ML-KEM draft, if we
are to publish it.

I want to assure you that people who prefer non-hybrids have done their 
analysis, not less rigorously than the proponents of the hybrid approach. Both 
groups are well-aware of their opponents’ arguments. The fact remains that 
these two groups disagree in their conclusions.

For how much longer will the opponents of non-hybrid (because the other side 
has nothing against hybrids — they simply want the freedom to exercise other 
options, based on their educated conclusion) beat this dead horse?
_______________________________________________
TLS mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to