Re: [Uta] Genart last call review of draft-ietf-uta-smtp-tlsrpt-17

[Jim Fenton <[e-mail]>]

On 3/26/18 5:40 PM, Brotman, Alexander wrote:
> So, we'd need to create a new Service Type (I see only  * and email currently), correct?  Could we alternately use the "n=" field and require that the note be set to "tlsrpt"?  If we do add this, and the receiving party is unable to find the appropriate s= or n= for the key, should they then discard the report?

In RFC 6376, the description of n= begins, "Notes that might be of
interest to a human (qp-section; OPTIONAL, default is empty). No
interpretation is made by any program." You're describing interpretation
that would be made by a program, so using n= would be the wrong approach.

Adding a service type involves putting an IANA consideration in the
draft saying that the specification requests that some value be added to
the DKIM Service Types registry. Shouldn't be hard to do, and it is the
Right Thing (IMO).

The s= for the key comes as part of the DNS response with the public
key, so there can't be a situation where you can retrieve the key but
not the service type. The default is "*", so if s= isn't there, you
should accept the report. If s=email (only) is there, yes, you should
discard the report, since eliminating that sort of abuse is the reason
for doing this in the first place.

-Jim