[saag] Re: Updating dkim standard to forbid rsa 1024 after deepmind paper on rsa ?

[Loganaden Velvindron <[e-mail]>]

On Sun, 8 Nov 2026 at 20:13, John Levine <[e-mail]> wrote:
>
> >>Im wondering if dkim standard should have a bis or a deprecation for
> >>rsa 1024 ?
>
> tl;dr  No.
>

> The DKIM WG is developing a follow-on to DKIM, currently called DKIM2.  It uses
> the same keys as DKIM and its predecessort Domain Keys.  In DKIM2, the currently
> optional Ed25519 signature is made mandatory and we'll add a PQC scheme.  Nobody
> is working on the existing DKIM spec which we expect to be obsolete in a few years.
>
How many implementations of DKIM2 are already available ?

> We are not the Network Police and telling DKIM users to stop using 1K keys will
> have zero efffect. (They don't use Ed25519 keys either, not even me and I wrote
> the RFC.) Mail operators who are paying attention are aware that 1K keys are too
> weak and have moved to 2K. Mail operators who are not paying attention will
> continue not to pay attention.
>
I agree with you. However, having a formal RFC would nudge those who are still
using 1K keys to become compliant right ?


> Before someone says stop accepting 1K signatures, nope. They're signing real
> mail, and the definite loss of valid mail with weak keys far outweighs the
> hypothetical threat that someone might forge a few short lived DKIM keys.
>
> R's,
> John
> _______________________________________________
> saag mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]