Thats about the number we maxed out on, but it wasn’t that formal of a test, and whats using webrtc, was just a stress test on conference to find breaking points. I’d like to see more info on the verto web socket breaking points if you could switch it to just audio and see if you can still see that break?
On Sep 29, 2016, at 11:18 PM, Chad Phillips <chad@apartmentlines.com> wrote:
Just finished a test where I got 61 simultaneous users into a FreeSWITCH videoconference via Chrome/mod_verto.
Here’s a link to a youtube video of a portion of the test: https://www.youtube.com/watch?v=c0kgb1zNxMU <https://www.youtube.com/watch?v=c0kgb1zNxMU>
Some stats for those interested:
FreeSWITCH box: 61 simultaneous WebRTC video feeds via Chrome/Verto Conference configured to minimize outbound encoding 32 core Xeon E5-2650, 32GB RAM Load average seemed to peak around 16
Test users: 10 Headless Ubuntu test servers, 20 cores / 8GB RAM each Each box running 6 Chrome instances via Selenium
The video quality was a tad blurry, not sure if that was something on my end or the muxing end of things. Also, I had to put the users in 12 at a time, as I’ve found that too many joining at once is causing websocket disconnects, not sure what the deal is there yet.
Curious if anyone else is trying crazy stuff like this, or if I’ve just set some sort of record. :P