Approximately how much bandwith is required per user-hour on a call?

I’m doing an evaluation about whether my company would like to consider hosting Jitsi on our AWS infrastructure for internal use, and the main consideration is cost. Are there numbers available about how many GB of bandwidth are used by Jitsi on average? If I have that I can do a back-of-napkin estimation about whether the cost is feasible for us.

Also, from this source: it looks like a single VPS or docker container should be able to handle several hundred simultaneous streams comfortably. Is that assumption correct?

This had been discussed several time in the comunity

I don’t really understand this math. Shouldn’t each user be 4Mbps up and 5.8 down, not 20 up and 5.8 down?

These calculations are from the servers point of you. As cost you are caclulating for server.

Ok so this 12-18 GB/Hour is a rough estimate for the total transfer cost for a 5 person conversation, is that correct? As in roughly $1-$2 total per hour, not per user per hour

1 Like

Top query - were there answers?