Re: SIP/PSTN interworking draft

[Scott Petrack <[e-mail]>]

It would be a very interesting thing to have a standard 
format for Calling Card Information, so that it could be 
included within, among other things a SIP payload. 
This would include the information on a "usual" calling
card, such as the carriers name, the owner's name, the
calling card number, perhaps a few bytes of special
restrictions. The whole thing would be protected via
either a password (analogous to the PIN) or via 
public-key. 

The point of all that is -- surely the correct format
has nothing to do with OOB dtmf. In fact, dtmf could
not be used, since the only way to tell in a stream
of random dtmf digits which is the calling card number,
which is the pin, etc., is to correlate the digits
with other real-time prompts. ("please enter your pin",
"please enter your calling card number"). Therefore, 
such dtmf information would need to be in band. 

Bottom line -- calling card info, yes. 
DTMF strings within SIP, no.


Final suggestion -- if there really is some 
need to send DTMF within SIP, I humbly suggest
that the dtmf be put into a MIME payload of type:
audio/telephone-events which is being defined
right now in AVT. The encoding of this MIME
type should be exactly -- the RTP dtmf payload
type. Why invent a new encoding? 

Scott



On Wed, 21 Jul 1999, Adam B. Roach wrote:

> 
> My interpretation of the feedback I had at last week's MMUSIC meeting 
> on draft-roach-mmusic-sip-pstn-require-header-00.txt is as follows:
> 
> 1. In-band DTMF transmission should be removed from the draft.
> 
> 2. Call-control services should be removed from the draft.
> 
> 3. Out-of-band DTMF transit is controversial.
> 
> My purpose in this message is to check if there are any counter-arguments
> to the first two points which anyone would like to present, and to
> begin a dialogue on the third point. 
> 
> To summarize what I beleive the main arguements for and against inclusion
> of OOB DTMF:
> 
> - Christian Huitema made the point that an interworking draft should
>   limit itself to those features absolutely necessary for signalling
>   interworking. Since DTMF isn't a necessary part of call setup and
>   teardown, it doesn't belong in this draft.
> 
> - Steve Donovan contends that, without having a requirement to send
>   DTMF information in a format understood by SIP nodes, the most
>   basic services (such as calling card services) will be impossible
>   to implement. PSTN interworking is of limited value if you can't
>   authenticate users.
> 
> I'm on the fence, myself. Both seem like valid arguments. I'd appreciate
> it if anyone who has a stake in this arena would weigh in on one side
> or the other. Thanks.
> 
> -- 
> Adam Roach, Ericsson Inc. |  Ph: [phone] | 1010 E. Arapaho, MS L-04
> [e-mail]   | Fax: [phone] | Richardson, TX 75081 USA
>