[Sipping] Re: Comment on draft-narayanan-sipping-aaa-diameter-00.txt

Cullen Jennings wrote:
> I am very happy to see some work in this area - thanks.

Thanks for your comments.

> I think a better solution would involve a network generated billing id. (ala
> DCS)

Generated by the proxy server that sends the DIAMETER messages?

> 
> Accounting-Input/Output-Octets/Packets serves no purpose I can imagine.

This was motivated by the need to deal with "user-to-user" information 
in SIP messages. There appear to be only two approaches:

- Try to prevent user-to-user signaling by stripping SIP signaling to 
keep SIP from becoming a free (3G) IM / SMS service. In my opinion, 
futile and counterproductive.

- Account for user signaling messages, so that users can be charged if 
they decide to embed their file transfers into headers of the SIP INVITE 
messages.

> 
> I think it would be good to look at the things people are currently
> collecting from SIP and H.323 calls in billing system. Figure out what they
> want to accomplish with this information then figure out a good way to do
> that.

Advice and insights on this would be much appreciated. We want to 
accomplish at least two things:

- gain insight into user SIP message behavior statistics
- session duration accounting

(I won't repeat the usual caveats of using SIP information for the latter.)

> 
> If you are defining some AVP, you might want to make the media ones specific
> to RTP instead of SIP so they can be used by H.323/MGCP/H.248.

We had discussed this and hope to get to that.

> 
> I'd give some thought to what you would need to bill for a call that got
> transferred and what is needed for a conference call.

I'm hoping that a "central mixer" conference call looks similar to a 
point-to-point call. Transfer and redirection is tricky, since there are 
many parties that could be charged, depending on the business model.

> 
> Hope that helps,
> 
> Cullen

_______________________________________________
Sipping mailing list  https://www1.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors@cs.columbia.edu for questions on current sip
Use sip@ietf.org for new developments of core SIP