RE: CBV

> From: [email protected]
> Sent: Friday, May 07, 2004 3:01 AM
> To: [email protected]
> Subject: RE: [srs-discuss] CBV
>
>

<...>

> Surely, at this point, a simpler and easier solution would be only to
> accept PGP signed messages from a verifiable source, and possibly include
> a set of permissible source addresses in the PKI infrastructure with each
> public key.
>
> That's what this starts to sound like.

PGP has always been on the table as a possibility.  Doing a body checksum
does move SES a bit closer to a public key system, but so far without the
PKI.  I don't particularly like this development, either, but I think we
should continue the discussion and see where it leads.

As far as PGP itself goes, the PKI is a fair amount of baggage and the trust
model is not easy to deal with for average users.  More importantly, I don't
think it solves the original problem we have set out for ourselves.  That
is, provide a reasonable amount of accountability for senders of email
without putting an unreasonable burden on either party, but particularly on
the recipients.  SPF, SRS and SES all attempt to validate MAIL FROM:.  PGP,
on the other hand, was meant to verify the RFC2822 headers, protect the
message body and allow for public key encryption.  It does nothing to
authenticate MAIL FROM:, which SPF correctly identified as the first
necessary step in providing accountability to senders.

Your suggestion of putting permitted sending machines into the PKI is very
interesting, but the nature of the PKI makes that information unreliable.
Unlike the DNS, anyone can put a cert into the PKI without the domain
owner's knowledge or permission.

--

Seth Goodman