using conference app in aws cluster of ant media server

[vardhaman surana]

Hi,

I will be using cluster architecture for our e-learning solution. I am coding according to the suggestions given in the following blog:

https://antmedia.io/how-to-create-a-webrtc-conference-room/

 

In this from what i understood, A webrtcAdaptor object will be created for a client, and then using the publish and play methods in that webrtcAdaptor object,  client can publish and play streams. But in the cluster architecture it is advised that publish request should go to origin servers and play request should go to edge servers.

So I just want to know that webrtcAdaptor.publish() and webrtcAdaptor.play() methods handle the origin and edge server thing internally or i will need a new webrtcAdaptor for each stream . i.e. in conference if i am playing two other streams and publishing mine then 3 webrtcAdaptor objects will be created ?

[Tahir]

Hi,

In Conference if you use the same webRTCAdaptor object for publish and play, it will publish and play from the same instance. In your example if you are playing 2 streams and publishing yours, you need two webRTCAdaptor object not three. One of publishing one for viewing. Also, if you enable sticky sessions it will go to the same instance even in that situation. 

Hope this may clarify things for you a bit. 

Thank you for using Ant Media.

[vardhaman surana]

thanks tahir, one more thing i wanted to know that in the cluster architecture after the peer connection has been made, the stream data flows through load balancer or not. Just need an estimate for pricing since aws charges load balancer prices for data flowing through it. 

Also can the peer connection be made to server even if does not have a public ip ?

[Tahir]

Hi,

If you are using webRTC and there is no specific rule or something it will not flow through the load balancer, it will go directly to the server. However, if you are using HLS it will flow through the load balancer.

Thanks.