Alistair Cunningham wrote: > > It's not quite the question you're asking, but I'll weigh in anyway. > When implementing this, please bear those writing billing engines in > mind. For example, imagine a call comes in from the PSTN to a toll-free > number (so the inbound call must be billed). The caller listens to an > IVR for 10 seconds, then talks to a person on a locally registered SIP > handset for a further 20 seconds. That person then blind transfers them > out to a cell phone (and the outbound call must be billed). They talk to > the cell phone user for 30 seconds, who uses the # key to attended > transfer them to a different cell phone (which is a separate billable > outbound call), with the first cell phone user talking to the second for > 15 seconds during the attended transfer. The original caller talks to > the second cell phone for 2 minutes, then hangs up. The inbound call > must be billed for 3:15, the first outbound for 0:45, and the second > outbound for 2:15 (ignoring ring times on the outbound calls). Of > course, you don't need to implement this logic in Asterisk, but please > keep the model of the channels consistent enough that someone writing a > billing engine can figure out exactly what call legs they need to bill, > and for how long, and when transfers occur which endpoint did the > transfer and how. AGI suffered from billable call legs appearing to hang > up when transferred (even though they stayed up), making it very > difficult to piece together exactly what billable calls occurred when. > It would be good to avoid this in ARI. From the perspective of the transfer events you'll get full details about everything involved. Who did the transfer, what type, the bridge involved, the channels transferred, where. The ARI application can interpret these and do what it wants with them. -- Joshua Colp Digium, Inc. | Senior Software Developer 445 Jan Davis Drive NW - Huntsville, AL 35806 - USA Check us out at: www.digium.com & www.asterisk.org
