Got things scaling well, few issues but due to more code required on my end. I was wondering though if there’s any noticeable ways to lessen CPU usage of the consumer/producer transports?
I was able to push 70 consumerTransports on one core which equated to 70 users viewing one stream or scaled up.
Yup, I’ve got my setup piped and sorted across an internal network but was trying to maybe pinch for more performance or switch hardware for this.
I saw there was a ~500 consumer ability but wasn’t sure if that was in reference to a single core in use, or what the configuration was. Single core on a first gen xeon is pushing 2 producers@1%; PipeTransports@0.3%; 2 consumers@1%.
Thank you for your time, spent a lot reading and just curious and if there’s a tip to play with these loads by adjusting bitrate or something, I’m all ears!
Mint! Thanks IBC, and thanks for that clarity @nazar-pc.
Been so busy storming the server API and scalability, test time is coming into play and want to be on the ball. Community is great and I’ve honestly learned lots with the well constructed documentations–cheers!
Do hope in time I can offer more than just “coder” insight, this is very powerful and I do intend to learn more.
Time to explore these results, and compare see what I need to do. Happy Easter.
May I ask which CPU was demo’d in the posted results and second (again sorry), is there any way to lessen the usage of the transports? My consumer takes uip a bit of load with having to re-produce the streams and consume them and would love to lessen this any way possible. Open to ideas/etc but to state I tried lower bitrate and no benefit. Are there any settings/techniques? Thanks again just wanting to solidify my answer so I can apply myself properly.
