We have deployed a UCCE 9.0 with CVP 9.0, using CUBES to connect to the outside world.
Because our client would like to use CVP survivability in the future we already enabled inactivity detection as follows:
gateway
media-inactivity-criteria rtplib
timer receive-rtcp 4
timer receive-rtp 1200
It turns out these settings have an influence on back office calls (not passing through the CC, i.e. direct inbound calls) as well.
Calls not answered within 20 seconds are cancelled by the CUBE.
According to Cisco TAC this behavior is by design, as the initial INVITE from the provider contains a DSP (early offer), which activates inactivity detection for that call.
I've been reading up a bit on inactivity detection, and I have the impression that it can be activated on specific applications or dial peers, i.o. on the CUBE level. I'm not a big CUBE/gateway expert, so I would appreciate if anyone can confirm this.
Or are there other ways to circumvent this unwanted behavior?
Thanks in advance!
Koen