[cabf_validation] OrganisationIdentifier mandated by ETSI TS 119 495

[Ryan Sleevi]

On Mon, Nov 5, 2018 at 10:40 AM Adriano Santoni via Validation <
validation at cabforum.org> wrote:

> Dimitris' description is correct.
>
> Just a remark on qualified certificates for electronic seals (a strange
> beast to all non-European CAs, I guess), since you mention them: they are
> not provided for as an alternative to QWACs - they are (will be) used for
> different purposes, that is for signing requests by the TPP (if so required
> by the ASPSP) at application level. But QWACs will always be required. The
> communication between the TPP and the ASPSP must always be secured by TLS
> 1.2 or higher, and in that context QWACs must be used. At least, this is
> my understanding, based on presentations and papers that I am aware of. The
> ETSI people involved in Open Banking standardization may correct me if I am
> wrong.
>
Again, the requirement to use a QWAC does not imply the requirement to use
a publicly trusted certificate.

For the same reason that the ecosystem, in a variety of sectors, has
recognized value in *not* using publicly trusted certificates, I think it's
extremely foolish to suggest that there is somehow a need that the
certificates MUST be recognized by browsers as well.

As we saw with the deprecation of internal server names, there's benefit to
private PKIs.
As we saw with the deprecation of RSA-1024 bit certificates, there's
benefit to private PKIs.
As we saw with the deprecation of SHA-1, there's benefit to private PKIs.
As we've seen in plenty of government PKIs for national ID, there's benefit
to private PKIs.

A number of members in this Forum happily and competently offer
privately-managed PKIs, ranging from everything such as employee identity
services to serving as management for internal devices and everything in
between. In the context of qualified certificate, you can provide both
QWACs and PSD2 "compatible" certificates without running afoul of the
Baseline Requirements, by ensuring these certificates are not publicly
trusted for TLS.

I know we've been going in circles on this topic for months, and that's
because we need to recognize the situation:
1) A number of CA/Browser Forum members participate within ETSI but did not
raise the concerns of BR incompatibility to the group
2) The ETSI ESI did not examine for incompatibilities with the CA/Browser
Forum Baseline Requirements until after it had ratified the document
3) There's no technical or policy reason to require the use of PTCs in the
context of PSD2
4) The objectives can fully be met, without any changes, through the use of
private PKIs, particularly since this primarily applies to inter-bank
communications

Just in the same way it would be foolish to suggest that, say, the ATM
network should use PTCs, or that of, say, the PKI used for drone management
or inter-car networking, there's no meaningful technical requirement to use
PTCs that are QWACs here. It does not provide any technical value, and just
introduces a host of greater risks to both parties - PTCs in the sense of
incompatible specifications, and Qualified PSD2 certificates in the similar
example of deprecations like SHA-1 or the improvement of validation methods.

Good software design is about recognizing when there are different,
incompatible threat models, and using the right tool to mitigate the risk.
I'm not saying PSD2 certificates are "less" secure, but I'm saying that the
needs of that ecosystem are functionally distinct from that of the TLS
ecosystem at large, and that's OK.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cabforum.org/pipermail/validation/attachments/20181105/97f08b58/attachment.html>