Melinda,
We have provided feedback on redaction earlier on this forum and also on
some other public forums.
We've heard loud and clear from customers in many product segments that
they need redaction for publicly-trusted certificates because
they value privacy.
We don't think redaction as currently specified in RFC 6962-BIS makes much
sense; it's too complicated and doesn't protect the privacy of domain
owners. If
un-redaction is possible by just throwing some cycles at it, then nothing
has really been redacted.
 

To summarize the position we've already stated: we feel that the issuing CA
must be approached by the domain owner to achieve disclosure of the full
contents of a redacted pre-certificate. Policies and procedures should be
crafted to
standardize that, but they shouldn't be tied to a discussion on redaction
because they are needed even when a pre-certificate is not redacted.
 
A redacted pre-certificate contains 95% of the information in the
certificate, and
fully demonstrates CA compliance to policies and requirements. A redacted
pre-certificate contains 100% of the information needed for a harmed party
to take
recourse (issuer, serial number, notBefore, owned domains in the redacted
CN
and DNS SANs). The recourse for a redacted pre-certificate is exactly the
same as the
recourse for an un-redacted pre-certificate: approach the CA, within the
terms of its
Relying Party Agreement, CP, and/or CPS and state a claim that a
certificate
has been issued in error and it encroaches on the rights and property of
the
complainant. The only action that can be taken by the issuer is to revoke
the certificate. Whether a browser can or cannot consume certificate status
information in a timely manner is out of the scope of this RFC.
 
We favor the full redaction as originally proposed in earlier versions of
6962-bis. Redaction blocks FQDN components that provide intelligence about
internal networks that was designed to be parsed and understood only by
employees and peer-to-peer systems privy to internal DNS. Redaction blocks
geographical data embedded in FQDNs; it blocks harvesting of high value
targets prior to attack. Some of our largest customers have had structured
naming conventions for many years; forcing them to use wildcards or
redesign
their DNS hierarchy to preserve privacy in an eTLD+2 policy world requires
massive network renaming far beyond the staffing levels of current day IT
organizations. The type of very large organization that has a long and
vastly deployed legacy of intelligent naming and needs redaction also has
strict IT policies that require certificates to be signed by public
audited CAs.
 
If redaction is completely removed from 6962-bis, we believe that it should
be moved to another RFC so that work can proceed.



- Sanjay

On 8/12/16, 5:17 PM, "Melinda Shore" <[email protected]> wrote:

>Hi, all:
>
>Since there was no comment at all on the proposal to retain
>name redaction, we appear to have complete agreement that
>it should go.  We'll go back into wglc when a new version
>is submitted.
>
>Melinda
>

_______________________________________________
Trans mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/trans

Reply via email to