Hi everyone,
I am currently working on optimizing a streaming application to improve user experience and load times. I’m looking for insights on best practices for handling high-volume media requests and maintaining stable performance.
Are there specific patterns or architectural approaches you recommend when dealing with real-time media tools? I’m also documenting my own findings and performance benchmarks over at cloudstream.pk if anyone wants to take a look and share feedback. Any documentation or personal experiences would be greatly appreciated.
Thanks in advance for the help!"
Hi, and welcome to the developer community.
You may look into binary transport (e.g. protobuf) - i wrote my own implementation of that so instead of sending json I am sending binary packages from client to server and vice versa.
And you may also look into gossip mesh. I am using that for a videocall system where I am running a capability check on the participants and then put them in different meshes depending on their capabilty. This way some with higher capability get better video quality and others get lower…
think of it like SFU (which does this automatically .. kind of.. )
And for videos of course use the right codec. You want to send a keyframe once and then only deltas (the stuff that has changed since last frame).
And there is also QUIC protocol that allows for session retention and it helps alot when you don’t have to do a tls handshake for every other connection.
And last but not least: don’t send what you don’t need. This sounds obvious but every letter, every number, everything that you send that is not needed adds up to slow down your stuff…
Need more?
