Real-time communication is a tricky business, and large scale video conferencing is especially so. There are many things that can be done to improve user experience but they get harder and harder as you go through the list. Bridge cascading, which we’ll explain shortly, is on the hardest end of this list!
Traditionally in a video calling all the callers on a particular video conference are anchored to the same bridge in one physical location. It’s hard to know where everyone will be calling from in advance, so most systems out there simply schedule the conference on a bridge that’s close to the conference creator. This sounds simple but can cause a lot of quality issues – imagine a call with 20 people where the sole Sydney participant happens to join first … that’s pretty much a recipe for a massively disappointing experience, since audio and video for 19 people would literally be going all through the planet with long delays for no good reason.
This is what bridge cascading solves. It ensures that every participant in a conference call will be connected to a bridge near them. A cascade of bridges is used to fan out streams to the right locations. To pull this off, the bridges must communicate among themselves to relay the appropriate streams around the world.
After much engineering, we are happy to announce we have added this feature to Jitsi Meet:
Try it on meet.jit.si right now!
UPDATE:
See a video overview of how this works here. Boris also did a write-up on this for webrtcHacks here. If you want even more, you can also see the academic paper he published on this topic here.
Love,
You personal Jitsi team!
