Iñaki Baz Castillo wrote:
> 2008/12/18 Paul Kyzivat <[email protected]>:
>> The problem here is that you have a device that is trying to charge, but
>> controls nothing to enforce its charging. If it wants to charge, then it
>> ought to have something (e.g. the media stream) that it can "turn off" when
>> the call terminates. Then there won't be fraud.
>>
>> If the proxy doesn't control the media path, then *what* is it charging for?
>> If it is providing no services after the call is set up, then it should only
>> charge for call setup, not call duration.
>>
>> I realize many deployed systems do this anyway - its just popular to charge
>> for things that cost you nothing - its a good way to make money.
>
> Well, you can be a SIP provider an also a client of your carriers
> (those who provide the gateways, the RTP endpoints in calls to PSTN).
> In this scenario you don't need to handle RTP in your SIP
> infrastructure, do you? But of course you need to charge your clients
> (you are a traffic reseller, even if the RTP goes directly from your
> clients to your carrier providers).
>
> IMHO, the only solution here (without handling RTP) is a B2BUA that
> does accounting and also SessionTimers in clients leg and carriers
> leg.
You can get the same effect with a call-stateful proxy that uses Session
Timer.
But if you don't control the media, that will not protect you against
someone intent on defrauding you.
I could create some custom UAs, that establish a media stream using the
normal sip means, but that continue to use it even after sending BYE.
The bottom line here is that if you don't control the media, then it is
not *your* network that the media is using, it is the *caller's* network
that the caller is using for media, and you have no justification to
charge for it.
Thanks,
Paul
_______________________________________________
Sip-implementors mailing list
[email protected]
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
