Another use case for multiple DKIM signatures is as a blanket verification
mechanism.

Yahoo's complaint feedback loop normally expects ownership verification
through a DNS TXT insert.

For entities with a large number of sending domains (like a CRM platform),
Yahoo accepts signatures from a previously agreed-upon domain/selector pair
in lieu of a DNS insert.

- Chris




On Wed, Nov 4, 2026 at 4:09 PM Randolf Richardson, Postmaster via mailop <
[email protected]> wrote:

> > On 04.11.26 11:21, A. Schulze via mailop wrote:
> > >accedently I found a message having 8 DKIM signatures. My DKIM
> > >verifier (OpenDKIM) wrote a log "ignoring 2 signatures". Challenge
> > >accepted! I also found "ignoring 53 signatures"! *)
> > >
> > >The number of messages having *many* DKIM signatures is very low. But
> > >the exist.
> > >And for that reason it's good to have a limit like this:
> > >
> https://manpages.debian.org/trixie/opendkim/opendkim.conf.5.en.html#MaximumSignaturesToVerify
> >
> > >This is also mentioned in RFC 6376, Section 4.2
> > >
> > >>To limit potential denial-of-service attacks, Verifiers MAY limit the
> total
> > >>number of signatures they will attempt to verify
> > >
> > >Have a minute to check that your DKIM validation implementation have
> > >limits implemtent and activated ...
> >
> > I just want to point out that the default for this is according to
> opendkim
> > manpage "3" which I believe should be enough most of the time.
> >
> > Also, the same doc indicates that signature for domain in From: is
> always
> > verified.
>
> To add to this good information (above) is that two common reasons
> for having multiple DKIM keys is:
>
> 1. to support re-signing existing DKIM keys in parallel with the old
> ones (usually only for a short during)
>
> 2. to support newer ED-signed DKIM keys in parallel with RSA-signed
> DKIM keys so as not to exclude older systems that don't support these
> newer algorithms
>
> RSA 2048-bit DKIM keys are the most common, although many systems are
> still using only RSA 1024-bit DKIM keys (these DKIM keys should be
> upgraded to at least 2048-bit as soon as possible).
>
> ED25519 and ED448 keys are newer, but not as widely supported,
> although I hope this will cease to be a problem soon.  While ED25519
> has the smallest key size, ED448 has a slightly larger key size with
> the benefit of better security (both are a lot smaller than the RSA
> keys, which is also much nicer for DNS record maintenance).
>
> I suspect that as more mail systems begin to transition to supporting
> ED-signed DKIM keys that the presence of multiple DKIM keys will also
> become more common as long as there is a significant quantity of mail
> SMTP servers that only support checking RSA-signed DKIM keys.
>
> The following Linux script I wrote generates all these keys, just in
> case anyone is interested in seeing the differences, and also
> possibly deploying these:
>
> #!/bin/bash
> # DKIM key generation script by Randolf Richardson
> # in Beautiful British Columbia, Canada
> export DATE=`date +'%Y%m%d'`
>
> # RSA (modern OpenSSL versions default to 2048-bit)
> echo openssl genpkey -algorithm rsa -out dkim_private_${DATE}rsa.pem
> echo openssl pkey -in dkim_private_${DATE}rsa.pem -pubout -out
> dkim_public_${DATE}rsa.pem
>
> # ED25519
> echo openssl genpkey -algorithm ed25519 -out
> dkim_private_${DATE}ed25519.pem
> echo openssl pkey -in dkim_private_${DATE}ed25519.pem -pubout -out
> dkim_public_${DATE}ed25519.pem
>
> # ED448
> echo openssl genpkey -algorithm ed448 -out dkim_private_${DATE}ed448.pem
> echo openssl pkey -in dkim_private_${DATE}ed448.pem -pubout -out
> dkim_public_${DATE}ed448.pem
>
> --
> Postmaster - [email protected]
> Randolf Richardson, CNA - [email protected]
> Inter-Corporate Computer & Network Services, Inc.
> Vancouver, Beautiful British Columbia, Canada
> https://www.inter-corporate.com/
>
>
> _______________________________________________
> mailop mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
>
_______________________________________________
mailop mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to