On Jun 21, 2010, at 2:32 AM, Magnus Burman wrote:
Hi Adrian,
I struggle to figure out how the MaxSessionTime for new calls are
calculated. As a simple example, have a first session ask for 3600s.
Then a second session does the same directly thereafter, with the
same rate (for simplicity), but credit is low and only 2000s is
allowed. How does call control figure out that both calls should end
after 2800 seconds?
This is done by recalculating based on the remaining balance a new max
session time for each individual session. The remaining balance is
redistributed in such way that all calls end up at the same time when
the balance reaches zero. So the Rating engine does not only return to
the call control the max session time for the current call but also
for previous calls if any are ongoing.
Check Prepaid documentation from CDRTool where this is explained in
detail.
Even more importantly to me is what happens when a third call is
entered into the mix. The credit is now at 0, is this call
established?
What I'm trying to achieve right here is handling multiple sessions
without blocking the credit. By having call control re-ask for
MaxSessionTime after a shorter time-period (60s), blocking will not
occur to such a large degree.
Again you are trying to solve a problem that does not exist.
Is this a bad way of doing it? I'm actually trying to integrate this
with a 3rd party rating engine that supports this schema as I
believe it's a good way of handling multiple sessions.
Read the documentation, it contains pseudo code for how to do it in
other rating engine app.
Cheers,
Magnus
2010/6/18 Adrian Georgescu <[email protected]>
The call control engine + rating engine already supports multiple
parallel prepaid sessions per user, all calls stop when the balance
reaches zero.
What do you try to achieve?
Adrian
On Jun 18, 2010, at 4:56 PM, Magnus Burman wrote:
Hi guys,
To get a more flexible credit control when running several
concurrent calls on the same user I'm trying to add some
functionality to the call control module.
Basically what I'm trying to do is set a low time limit, such as 60
seconds, and when the timer runs out send a new getCallLimit, until
the call ends either by a return of 0 or by hangup.
I've been trying to do this by adding a new getCallLimit in the
Call.__expire method in sip.py.
def __expire(self):
rating = RatingEngineConnections.getConnection(self)
rating
.getCallLimit
(self).addCallbacks(callback=self._reinit_simple_calllimit,
errback=self._start_error)
"""
self.expired = True
self.application.clean_call(self.callid)
self.end(reason='call control', sendbye=True)
"""
In _reinit_sample_calllimit I mimic _start_finish_calllimit in a
lot of ways, trying different timer combinations. I've tried with
cancel followed by a new setup, I've tried delay, I've tried reset,
I've tried manually doing a new ReactorTimer and all of the above!
I get the MaxSessionTime call to work, but I fail with the timer
somehow. __expire is never called again after the first time, and
the call is never interrupted.
Anyone with tips? At this point I'm about to dive into the twisted
framework and reactors, but any hints are much appreciated!
Cheers,
Magnus
_______________________________________________
Users mailing list
[email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
_______________________________________________
Users mailing list
[email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
_______________________________________________
Users mailing list
[email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
_______________________________________________
Users mailing list
[email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
