This question originally came up in our Slack community and the thread has been consolidated here for long-term reference.
For multi-region deployments, do we need a global Redis instance?
I was reading the scaling WebRTC blog post and it seems that every datacenter/region has its own Redis (Message Bus), with inter-datacenter communication happening via Media Relay.
The mesh system described in that blog post is specific to LiveKit Cloud, our commercial product.
LiveKit’s open source media server supports multi-region deployments, but those regions do not communicate with one another out of the box. Each region operates independently with its own Redis instance.
The media relay logic for cross-region communication is not available in the self-hosted version.