Mixing bridge vs. holding with ARI


Do you know if mixing bridge users significantly more resource then holding bridge with ARI (version 16)?
We have an application which broadcasts about 100 simultaniesly calls to +1000 listeners. We use holding bridge with the lecturer as the announcer. Lately it causes the lecturer to disconnect - probably because there is no RTP going back on the lecturer line, it makes our carrier to assume the call dropped. Mixing bridge does not have this problem, but we want to be sure we will not kill our server with so huge conferences. Of course we will mute all listeners channels.

The mixing bridge would use more resources, as it’s producing a stream, transcoding it out, etc. The specific impact on your system would depend on the system itself.

My main concern is whether it will cosume more resources per listener, or only per conference. In other words, will the increse of resources used by mixing compare to holding is in parallel to the number of channles on the conference?

Both. The more channels the more resources, both per-channel and for the conference itself. That being said - comparing this specific scenario is not something anyone has done to the best of my knowledge.

I hate to be the first trying…
So, another idea we had is to create snoop channel for every lecturer, and puting this channel as the announcer. This way, the lecturer channel still gets RTP back. Can you estimate which solution (mixing or snoop channels) will be less resource heavy?

Mixing probably less.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.