Daniel Van Geest
<[email protected]> writes:

> Some thought on whether the small or fast versions would be more
> appropriate for TLS could help then the codepoints to a reasonable
> number.

That is a good observation about SLH-DSA generally, and I agree some
guidance on how to pick the right variant can help people deploying it.

Below is my personal decision flow chart for picking SLH-DSA variant.
Can others share their thinking of SLH-DSA variant choices?

Is SHAKE available in your environment and its performance is
acceptable?

  YES: go with SHAKE variants
  NO: go with SHA2 variants

Is ~20x faster/cheaper signing more important than making verification
~3x slower/expensive?

  YES: go with fast variants
  NO: go with slow variants

Do you consider AES-128 security levels acceptable?

  YES: go with 128 variants
  NO: chose 256 variants

There are exceptions to those simple rules.  I've experienced a
situation where _both_ SHA2 and SHAKE were available, which I suppose is
rather common, but the significant performance increase with SHA2 made
it possible to get a ~7x improvement of verification speeds over any
slow/fast combination with SHAKE.  In that situation the final choice
was between SHAKE-fast (because SHAKE-slow signing would be a
bottleneck) and SHA2-slow (which had acceptable signing performance),
and for me SHA2-slow won because the ~7x verification speed improvement
became the deciding factor.

/Simon

Attachment: signature.asc
Description: PGP signature

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

Reply via email to