Re: [Sip] New I-D on RFC4474 and phone numbers

[Jonathan Rosenberg <[e-mail]>]

Elwell, John wrote:
> Yes, I think this is the right direction. I would not want to prevent
> investigations on how to do it better, e.g., based on the ENUM tree in
> some way. But what you propose is a practical step that gives some value
> in the short term, both in terms of anti-spam and DTLS-SRTP.
> 
> Perhaps we also need to do even more to promote email-style SIP URIs,

well, having tried to do that for something like ten years now, and 
having STILL seen the industry primarily doing pstn replacement which 
requires phone numbers, I am somewhat skeptical that there is anything 
IETF can do here.

> alongside Tel URIs for use when interworking with PSTN. This does not
> necessarily prevent the use of numeric values in the user part of
> email-style URIs (without user=phone), thereby making it easier to
> correlate the a SIP URI with the telephone number that would need to be
> used to contact the user from the PSTN. Perhaps it would be beneficial
> to provide both an email-style SIP URI and a Tel URI in From, so that
> the Tel URI can be used if the call breaks out to PSTN (where end-to-end
> security is not achievable) and the SIP URI can be used in other cases.

Now that is a topic worth discussing.

 From can only have a single URI. Interestingly PAI does allow more; you 
can assert both a tel and a SIP URI. What would be really nice is to be 
able to assert both the phone number and the username in the From and 
have the domain covered by the 4474 signature. As my draft points out, 
there is some small value in the 4474 signature there.

One way to do this for example would be something horrible like this:

From: sip:jdrosen*[e-mail];user=composite

where "composite" means that the user part is a series of names within 
the domain, each one separated by "*".

Unfortunately that would work much better in theory than in practice 
since I think most implementations ignore the user attribute of the URI 
and would not know that they shouldn't render this weird looking user part.

Probably the better solution is some hybrid where PAI is used for the 
phone numbers and 4474 for the user@domain identifier. Thats another 
thing useful we could do, is to try and clarify how PAI and 4474 work in 
concert. Right now that is clear as mud.

-Jonathan R.


-- 
Jonathan D. Rosenberg, Ph.D.                   499 Thornall St.
Cisco Fellow                                   Edison, NJ 08837
Cisco, Voice Technology Group
[e-mail]
http://www.jdrosen.net                         PHONE: [phone]
http://www.cisco.com
_______________________________________________
Sip mailing list  http://www.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use [e-mail] for questions on current sip
Use [e-mail] for new developments on the application of sip