To close the loop on this, I've proposed a PR to add this to the threat
model and security considerations explicitly:

https://github.com/ietf-wg-dnsop/draft-ietf-dnsop-domain-verification-techniques/pull/228/changes

On Wed, Oct 14, 2026 at 5:47 AM <lirx25=
[email protected]> wrote:

>
> Thank you for sharing your thoughts. I appreciate your perspective and
> understand the practical concerns you raised. I agree that using unique
> random validation tokens is an important best practice that should be
> encouraged across applications.
>
> AliasLeap could serve as a useful real-world example illustrating the
> potential security consequences of not adopting this practice. In
> particular, the risks are not limited to dangling DNS records but can also
> arise from CNAME following in otherwise legitimate DNS configurations.
>
> I believe briefly mentioning this scenario could further reinforce the
> importance of the draft's existing recommendations, without requiring any
> substantial changes to its scope or direction.
>
> Thank you again for considering my suggestions and for your efforts on
> this draft.
>
> Best regards,
> Ruixuan
>
>
> > -----原始邮件-----
> > 发件人: "Paul Wouters" <[email protected]>
> > 发送时间:2026-10-14 05:38:09 (星期三)
> > 收件人: [email protected]
> > 抄送: "Erik Nygren" <[email protected]>,
> [email protected], [email protected],
> [email protected]
> > 主题: Re: [DNSOP] Re: Security consideration for
> draft-ietf-dnsop-domain-verification-techniques
> >
> > On Tue, 13 Oct 2026, [email protected]
> wrote:
> >
> > > I agree with the draft's recommendation to use unique validation
> tokens, as this is an effective mitigation against the AliasLeap attack. I
> do not think the current
> > > recommendations are incorrect, but they may overlook some security
> risks. Therefore, I believe it would be reasonable to discuss the risks
> highlighted by AliasLeap
> > > in the Threat Model, Security Considerations, or the CNAME usage
> section.
> >
> > The document is mostly about how to do things right, not on explaining
> > the many wrong things one can do. The document aims at a concrete BCP,
> > and this is different from a whitepaper about a found security issue.
> >
> > > For these reasons, I believe that prohibiting CNAME following during
> DCV is a necessary security measure.
> >
> > That is going to be very very difficult, as most people who implement
> > DCV are application developers, and not DNS engineers/experts. They
> > will just use an existing DNS server or library and that in turn
> > will just follow CNAMEs. Asking them to develop some backend code
> > that has custom DNS processing will be very difficult and also
> > error prone.
> >
> > Therefor, randomization remains an essential property that developers
> > should not skip.
> >
> > Paul
>
_______________________________________________
DNSOP mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to