M2M call with no voice

This is merely a historical archive of years 2008-2021, before the migration to mailman3.

A maintained and still updated list archive can be found at https://lists.osmocom.org/hyperkitty/list/OpenBSC@lists.osmocom.org/.

Holger Freyther holger at freyther.de
Thu Dec 8 12:02:27 UTC 2016


> On 08 Dec 2016, at 11:29, Kristian Martens <kristian.martens at ng4t.com> wrote:
> 
> Hello,

Hi!


> I got a mobile to mobile call working for signalling only. Two RTP
> channels are being established successfully. But there are only a few
> pattern sent over these channels (1 byte 0x23 as UDP payload). Is there
> a setting I am missing in the configuration files? Please find attached
> a .zip file with pcap and cfg files.


I didn't look at the trace yet but have you seen my comment about the
SDP file?

<000b> mgcp_protocol.c:830 Got media info via SDP: port 5502, payload 112 (GSM), duration 20, addr 192.168.17.100
<000b> mgcp_transcode.c:170 Checking transcoding: GSM (3) -> GSM (112)

Why would one use a dynamic payload type for GSM FR1? I am not sure if
the mgcp_mgw is optimized to avoid doing transcoding if input and output
codec are the same but use different payload types. I would think it is.

The 0x23 UDP payload is to establish a NAT state to be able to send
traffic (e.g. ringtone) to a call for a BSC behind a NAT (all traffic will
go from network->NAT/Firewall->MGCP MGW) and the only way to know the
port is to have the BSC send some data.

So in your trace I don't see:

* UDP traffic from BTS to bsc_mgcp_mgw
* UDP traffic from bsc_mgcp_mgw to 192.168.17.100:5500
* UDP traffic from 192.168.17.100:5500 to ...

you get the point. Maybe start looking for where the BTS is sending the
traffic to and take it from there?



holger


More information about the OpenBSC mailing list