Scalability - Handling multiple concurrent sessions in Broadcasting mode

Issues with deployment
1

i was looking the section " How many users can an OpenVidu Pro cluster handle " https://docs.openvidu.io/en/2.13.0/openvidu-pro/scalability/

and was able to relate with one of my below use case

if i have deployed Pro cluster on AWS with 1 Media Node -
Type - c5.xlarge
Cores - 4
Memory - 8GB
Number of Publisher - 1 (Broadcaster publishing Audio and Video)
Number of Subscribers - 250 (Only receiving Audio and Video)
Stream Type - Audio + Video
Recording - Yes

I hope the above configuration will handle the specified load.

Now if a new session gets started at the same time i.e another Broadcaster created a new session and started broadcasting with the same numbers of subscribers ( 250 subscribers in this new session/room ). How this will be handled, will a new Media node will be created and load will be handled or do i need to create another Pro cluster with 1 Media node (one another OpenVidu server pro node and 1 Media node)?

If i need to create another Openvidu cluster then it means for every new session i will have to configure a new Pro Cluster (Considering 1 Broadcaster and 250 subscribers). If this is the case then how can we configure this Pro cluster on the fly?

2

Scalability in OpenVidu works in the following way:

  • Sessions are created in the less loaded node
  • Nodes can be added dynamically using a REST API
  • 2.14 doesn’t automatically add nodes on demand (2.15 will do. To be released the next week).
  • If you have 1 media node with a session consuming almost all CPU, create a new media node (with the REST API) so the new session will be created in the new (empty) media node.
  • You don’t need a new OpenVidu PRO cluster, just add a new media node to existing one.
1 Like