[cabfpub] Revisiting CAA

Thu May 1 22:59:32 UTC 2014

Thanks Rick.

I have no particular preferences regarding that last sentence - but agree,
it sounds good. The first sentence makes it clear in (iii) that it has to
be documented, which is the main goal.

The only modification I would suggest would be that you remove "IP
Addresses" from (i), since you only have CAA for Domain Names (as mentioned
in the text you quoted from 6844)

On Thu, May 1, 2014 at 3:26 PM, Rick Andrews <Rick_Andrews at symantec.com>wrote:

>  I’m attaching Phillip’s original proposal for CAA and Jeremy’s
> suggestion for enhancement. Here’s my proposal.
>
>
> Add to Section 4 Definitions, new item:
>
> *CAA*: From RFC 6844 (http:tools.ietf.org/html/rfc6844): “The
> Certification Authority Authorization (CAA) DNS Resource Record allows a
> DNS domain name holder to specify the Certification Authorities (CAs)
> authorized to issue certificates for that domain. Publication of CAA
> Resource Records allows a public Certification Authority to implement
> additional controls to reduce the risk of unintended certificate mis-issue.”
>
> Add to Section 7.1.2 Certificate Warranties, new item:
>
>         9. *CAA*: That, at the time of issuance, the CA (i) implemented a
> procedure for consideration of CAA records for each Domain Name(s) and IP
> address(es) listed in the Certificate’s subject field and subjectAltName
> extension; (ii) followed the procedure when issuing the Certificate; and
> (iii) accurately described the procedure in the CA’s Certificate Policy
> and/or Certification Practice Statement. It is permissible (although not
> desirable) for the CA to ignore CAA records completely, as long as that
> “procedure” is documented in the CA’s Certificate Policy and/or
> Certification Practice Statement. If the CA’s Certificate Policy and/or
> Certification Practice Statement is based on RFC 3647, the statement
> describing the CA’s CAA procedure SHOULD appear in Section 4.4.2.
> Certificate Application Processing.
>
> (I defer to Tom and Ryan S on that last sentence. You read many more CPs
> and CPSs than I do.
>
> -Rick
>
>
>
> ---------- Forwarded message ----------
> From: Jeremy Rowley <jeremy.rowley at digicert.com>
> To: 'Phillip' <philliph at comodo.com>, "public at cabforum.org" <
> public at cabforum.org>
> Cc:
> Date: Fri, 7 Jun 2013 14:32:44 -0700
> Subject: Re: [cabfpub] CAA Proposal
>
> Although I am starting to like the concept of CAA, I think this is an
> improper way to implement a CAA requirement.  If a CA already has rigorous
> validation practices and can accurately identify the request as originating
> from the proper entity, I’m not sure that additional checks are necessary.
>
>
>
> If we plan to implement CAA in the Forum, we should develop a discernible
> standard that can be used to measure compliance.  In fact, perhaps the RFC
> should be revised prior to the Forum’s adoption to identify what additional
> verification requirements should be considered necessary before issuance of
> a certificate.  That way the Forum has a basis for setting the additional
> checks and CAs will have a better understanding of how to comply.
>
>
>
> Jeremy
>
>
>
> *From:* public-bounces at cabforum.org [mailto:public-bounces at cabforum.org] *On
> Behalf Of *Phillip
> *Sent:* Friday, June 07, 2013 10:10 AM
> *To:* public at cabforum.org
> *Subject:* [cabfpub] CAA Proposal
>
>
>
> Following up on the CAA threads, I would like to propose the following
> (subject to discussion):
>
>
>
> 1) CABForum endorse the publication of CAA records by domain name owners
> to mitigate the risk of issue of certificates in response to an
> unauthorized or fraudulent request.
>
>
>
> 2) The Basic requirements be updated to add a requirement that CAs state
> their policy for use of CAA records in their CPS.
>
>
>
> "A CA MUST state its policy for processing CAA records as defined in RFC
> 6844"
>
>
>
>
>
> Rationale:
>
>
>
> http://tools.ietf.org/html/rfc6844
>
>
>
> To be compliant with the RFC, a CA MUST comply with the requirements of
> section 4:
>
>
>
> Before issuing a certificate, a compliant CA MUST check for
>
>    publication of a relevant CAA Resource Record set.  If such a record
>
>    set exists, a CA MUST NOT issue a certificate unless the CA
>
>    determines that either (1) the certificate request is consistent with
>
>    the applicable CAA Resource Record set or (2) an exception specified
>
>    in the relevant Certificate Policy or Certification Practices
>
>    Statement applies.
>
>
>
> A CA can be minimally compliant with the specification by simply
> publishing a statement that says that they retrieve and process CAA records
> for each request and then grant an automatic exception in every case.
>
>
>
> This is deliberate because there is a peculiar edge case in which the
> Domain Name owner does not control their DNS publication infrastructure and
> the party that does inserts a spurious CAA record to limit competition. It
> also avoided the need for theological debates on what is and is not a
> public delegation point.
>
>
>
> The point of CAA is to benefit CAs by reducing the cost of detecting
> potential fraudulent applications and mitigating the risk of issuing a
> certificate. But as with any other validation check, the response to a
> request that is non-consistent is not going to be to kick the request back
> to manual processing. There is going to be a person in the loop making
> enquiries. Either the CAA record is spurious and the CA wants to get it
> changed so that they can take the business or they have just detected an
> unauthorized request which they are going to want to look at an analyze and
> study.
>
>
>
> A CA could write a CPS statement that says they look at CAA records and
> then ignore them completely but that would not look good. I think it rather
> more likely that it would say something like they have some sort of process
> for determining that CAA records do not represent the intention of the
> Domain Owner and publish a list of domains they will ignore CAA records
> from. This might include top-level domains like .com etc. But the fact that
> CAs have the option of ignoring the CAA records is probably sufficient to
> deter an attack.
>
> _______________________________________________
> Public mailing list
> Public at cabforum.org
> https://cabforum.org/mailman/listinfo/public
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cabforum.org/pipermail/public/attachments/20140501/4a820f53/attachment-0003.html>