[cabfpub] EV Wildcards

[Ryan Sleevi]

Dean,

As we discussed on the past week's call, I think it's very important to
have the discussion first about what the information in the certificate is
supposed to represent, before we can have any truly meaningful discussion
about EV wildcards.

For example, we discussed at length at the call how a cloud provider could,
without any issues with the existing EV Guidelines, get an EV certificate
for each of their customers. By very nature of the the EV Gs, each of those
certificates would indicate the cloud host, rather than the actual operator
of the domain. Concretely, it would be entirely valid to issue an EV cert
to https://example.appspot.com that warranted itself as Google, even though
"example.appspot.com" might be run by Example Corp.

Unless and until we can address this, it's probably not going to be useful
to phrase the discussion as EV wildcards, but as what EV is to represent.
This is, notably, the same concern that some have raised with OV
information - for example, for cloud providers such as CloudFlare, should
the organization information represent who is terminating the TLS
connection (CloudFlare), who operationally controls the domain
(CloudFlare), or who technically administers the domain (the customer of
CloudFlare and ultimate operator of the site being accessed)

I think if we can establish these principals and reach some agreement, then
we can have meaningful discussions about both the validation methods of
certificates (since whose information are you validating?), and then also
have discussions about the nature of wildcards (since they may span
multiple domains)

On Thu, May 21, 2015 at 1:31 PM, Dean Coclin <Dean_Coclin at symantec.com>
wrote:

> Continuing the discussion we had at this past week’s CA/B Forum
> teleconference, we agree with those that were in favor of wildcard
> certificates for EV. The examples cited below exist today with DV wildcards
> which provide no authentication of the business. At least with EV, you have
> a strong level of business authentication and if the owner of an EV
> certificate were to mount a phishing attack, they can be clearly identified
> and disciplined. The same is not the case for DV. Hence it makes more sense
> to *only* allow wildcards for EV (or OV) rather than DV but clearly there
> was opposition to deprecating this for DV on the call (so I’ll leave that
> as a separate item for now).
>
>
>
> We have this as a discussion item for the face to face meeting along with
> validity periods although they are not necessarily tied together. I’ll try
> to prepare a pro/con sheet to guide the discussion so please continue to
> send replies.
>
>
>
> Thanks
> Dean
>
>
>
> *From:* public-bounces at cabforum.org [mailto:public-bounces at cabforum.org] *On
> Behalf Of *kirk_hall at trendmicro.com
> *Sent:* Thursday, March 26, 2015 12:03 PM
> *To:* Dean Coclin; CABFPub (public at cabforum.org)
> *Subject:* Re: [cabfpub] EV Wildcards
>
>
>
> Dean – while we are not passionate about this, the point we were trying to
> make is that with an EV certificate, users see the EV “green bar” and other
> information that is meant to enhance trust in the website.  Users also see
> the full FQDN displayed in the green bar *[high-risk].example.com
> <http://example.com>*, although most browsers appear to give prominence
> to the SLDN.  Who knows what users really think, but this seems to give
> legitimacy to the idea that for an EV cert for *facebook.example.com
> <http://facebook.example.com>*, the site has something to do with
> “Facebook” which it does not.
>
>
>
> Again, EV vetting requires the CA to screen for High Risk Certificate
> Requests, which we have always interpreted as applying to the entire FQDN,
> and we assume many or most CAs would not issue an EV cert for *facebook.example.com
> <http://facebook.example.com>*.  In contract, if we allow wildcard EV
> certs, there would be no way the CA could perform a screening for High Risk
> Certificate Requests above the SLDN, and so a potentially misleading name
> like *facebook.example.com <http://facebook.example.com>* could be
> displayed in the EV green bar.  That’s the situation that concerns us.
>
>
>
> Of course, BR 11.5 on screening for High Risk Certificate Requests
> (including screening for “names at higher risk for phishing or other
> fraudulent usage“ and “names that the CA identifies using its own
> risk-mitigation criteria“) already applies to ALL cert types, so I’m
> assuming a request for something like *facebook.example.com
> <http://facebook.example.com>* in a DV or OV cert would trigger some sort
> of extra screening today for most CAs (it does for us – we won’t issue that
> cert).  But there is less potential harm from wildcard DV and OV certs used
> to secure *facebook.example.com <http://facebook.example.com>* because
> the browser UI for users in DV and OV does not highlight the FQDN as
> prominently as for EV.  That’s our thinking.
>
>
>
> *From:* Dean Coclin [mailto:Dean_Coclin at symantec.com
> <Dean_Coclin at symantec.com>]
> *Sent:* Thursday, March 26, 2015 8:47 AM
> *To:* Kirk Hall (RD-US); CABFPub (public at cabforum.org)
> *Subject:* RE: EV Wildcards
>
>
>
> We disagree with this line of thinking. Today someone can pay a few
> dollars and secure *.example.com, where the result is [high-risk].
> example.com with the most limited form of authentication. However a
> legitimate organization that successfully passes EV verification cannot
> order that same certificate. This makes zero sense — in fact, since the
> concern is with the exploit, this logic means that wildcards would be
> forced to the least authenticated customers.  Hence we would support
> wildcards for EV certs.
>
>
>
> If we were going to change anything else (which is a completely separate
> discussion), it would be to put a minimum authentication requirement on the
> issuance of wildcards.
>
>
>
>
>
> Dean
>
>
>
> *From:* public-bounces at cabforum.org [mailto:public-bounces at cabforum.org
> <public-bounces at cabforum.org>] *On Behalf Of *kirk_hall at trendmicro.com
> *Sent:* Wednesday, March 25, 2015 9:05 PM
> *To:* CABFPub (public at cabforum.org)
> *Subject:* [cabfpub] EV Wildcards
>
>
>
> All the arguments on the question of allowing EV Wildcard certs are well
> considered and valid on both sides.  Chris Bailey and I come down on the
> side of “no EV wildcard certs” for the following reason.
>
>
>
> It’s true that *.example.com could “hide” the use for *facebook.example.com
> <http://facebook.example.com>*.  It’s also true that the owner of
> example.com today could ask for an EV cert for *facebook.example.com
> <http://facebook.example.com>*, and if the cert is issued, it’s “no
> different” from using a wildcard cert *.example.com.
>
>
>
> However, there is one important difference.  BR 11.5 (High Risk Requests)
> and the related EV Guideline 11.12.1 require the following:
>
>
>
> *BR 11.5 High Risk Requests*
>
> The CA SHALL develop, maintain, and implement documented procedures that
> identify and require *additional verification activity* for High Risk
> Certificate Requests prior to the Certificate’s approval, as reasonably
> necessary to ensure that such requests are *properly verified* under
> these Requirements.
>
>
>
> *High Risk Certificate Request: *A Request that the CA flags for
> additional scrutiny by reference to internal criteria and databases
> maintained by the CA, *which may include names at higher risk for
> phishing or other fraudulent usage*, names contained in previously
> rejected certificate requests or revoked Certificates, names listed on the
> Miller Smiles phishing list or the Google Safe Browsing list, *or names
> that the CA identifies using its own risk-mitigation criteria*.
>
>
>
> Among other things, we interpret that to require CAs to scan FQDNs for
> “names at higher risk for phishing or other fraudulent usage” at every
> level in the FQDN, and as a matter of policy, we generally won’t issue a
> cert for *facebook.example.com <http://facebook.example.com>* unless the
> customer can show us it has Facebook’s permission.  The same is true for a
> long list of other high risk names, and we apply the scan to all FQDNs in
> the SANs field as well.
>
>
>
> So this means that under our policy and interpretation a customer could
> not get an EV cert from us for *[high risk name].example.com
> <http://example.com>*, which helps cut down on the likelihood of fraud or
> imitation.  An EV cert for *.example.com, on the other hand, could be
> used to secure the same high risk name FQDN.
>
>
>
> We recognize that other CAs may not have a policy as restrictive as ours
> for EV certs, but if another CA issues an EV cert for *facebook.example.com
> <http://facebook.example.com>* and it’s used for fraud or phishing –
> presumably that CA will get very adverse publicity, and will have some
> explaining to do to the public.  That is likely to keep the number of such
> high risk name EV certs to a minimum.  In contrast, no scanning or review
> will happen for EV wildcard certs.
>
>
>
> So in that sense, there is a difference, and we think wildcard certs
> should *not* be issued for EV – prohibiting EV wildcard certs makes CAs a
> bit more responsible, in our opinion.
>
>
>
> *Kirk R. Hall*
>
> Operations Director, Trust Services
>
> Trend Micro
>
> +1.503.753.3088
>
>
>
>
>
> TREND MICRO EMAIL NOTICE
>
> The information contained in this email and any attachments is confidential
>
> and may be subject to copyright or other intellectual property protection.
>
> If you are not the intended recipient, you are not authorized to use or
>
> disclose this information, and we request that you notify us by reply mail or
>
> telephone and delete the original message from your mail system.
>
>
>
>
>
> TREND MICRO EMAIL NOTICE
>
> The information contained in this email and any attachments is confidential
>
> and may be subject to copyright or other intellectual property protection.
>
> If you are not the intended recipient, you are not authorized to use or
>
> disclose this information, and we request that you notify us by reply mail or
>
> telephone and delete the original message from your mail system.
>
>
>
> _______________________________________________
> Public mailing list
> Public at cabforum.org
> https://cabforum.org/mailman/listinfo/public
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://cabforum.org/pipermail/public/attachments/20150521/03b3dabe/attachment-0001.html