> -----Original Message-----
> From: Dean Willis [mailto:[EMAIL PROTECTED]
>
> You MUST sign in order to be able to use DTLS-SRTP effectively. This
> means that even PSTN-grade identities have to be signed. So we we either
> have to change DTLS-SRTP to not have this dependency, or we have to find
> a way to mark the PSTN-grade ones appropriately so that they are not
> taken seriously as a strong identity.

No, you don't have to rfc4474 sign dtls-srtp fingerprint attributes.  Rfc4474 
signing them prevents SIP devices between the signer and the verifier from 
modifying the attribute to become a dtls-srtp man-in-the-middle.  If a MitM can 
re-sign without the verifier detecting it, then rfc4474 signing is pointless, 
even for the DTLS-SRTP fingerprint attribute.

And that is the problem with e164.  If foo.com can receive sip:[EMAIL 
PROTECTED], and change it to sip:[EMAIL PROTECTED], and the UAS only displays 
"+12128675309" or the human only cares about that part because they think it's 
a phone number, then there's little point in signing.


> Adam's approach solves this better than us1ing PAID, as 1) we could NOT
> have DTLS-SRTP (a standareds-track doc) with a normative reference to
> RFC 3325 (an informational track doc). Further, the same problem applies
> -- even with RFC 3325, it would be good to differentiate strong
> identities from PSTN-grade identities.

You wouldn't need to reference rfc3325 in dtls-srtp-framework (and it's the 
framework this is about, btw, not dtls-srtp).  PAID has nothing to do with the 
fingerprint - it has to do with identity.  My point was you already do have 
PSTN-grade identities: PAID is it.

-hadriel
_______________________________________________
Sip mailing list  https://www.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use [EMAIL PROTECTED] for questions on current sip
Use [EMAIL PROTECTED] for new developments on the application of sip

Reply via email to