I'm currently configuring my two (2) servers to live-repeater origin/edge configuration
Server_1 has origin and edge applications.
Server_2 has edge application.
Since I'm using medium-sized standard EC2 instances, I'm trying to limit the CPU utilization by minimizing the number of transcodes, as well as the number of connections, to the origin server.
My plan would be like this:
Server_1 - I will limit the number of connections to the liveorigin application by using the module that limits the connections to an application. I will also limit the number of connections to the liveedge application by using the module that limits the connection to an edge server so that most of the connections will be redirected to Server_2 liveedge application.
Server_2 - The number of connections to the liveedge application will be several times more than that of Server_1.
I have tested this and this setup worked perfectly.
However, my concern is the transcoding process. I have read that the transcoder should be enabled at the edge application. That means, Server_2 will have a very high CPU utilization due to transcoding and high number of connections. But I could simply interchange the number of connections limits to the two liveedge applications so that CPU utilization of Server_2 will be lessened. So there'll be no problem about that.
But I have also read that the stream-type at the edge applications should be set to 'live' instead of 'liverepeater-edge' and that stream names could be simplified by using '.stream' file.
Now my questions are the following:
1. Do I have to set the LiveStreamPacketizers to xxxpacketizer,yyypacketizer,zzzpacketizer instead of xxxrepeater,yyyrepeater,zzzrepeater since I changed the stream type from 'liverepeater-edge' to 'live'?
2. Since I'm using an MPEG-TS-based encoder, do I have to use the Stream Manager to start the stream? If so, should I use the liveedge application to start the stream? What type of MediaCaster should I tick and what would be the structure of the stream name that I shall use?
By the way, I have created a 'mpegts.stream' file, whose content is udp://0.0.0.0:10000', in the /content folder since I'm using an MPEG-TS-based encoder. How could I use that if the stream name at the edge should look like this: rtmp://[live-origin-IP-address]/[liveorigin-app]/_definst_/stream-name.
Furthermore, I've also implemented 'load-balancing' (well, not really since the number of connections for each edge server are limited by 5:1 ratio) and I use 'Get least loaded server using Netconnection redirect' to get the least-loaded server.
Server_1: LoadBalancer Listener and LoadBalancer Sender
Server_2: LoadBalancer Sender
My player pulls the content using rtmp://[origin-server-IP-address]/[redirect-app]/stream-name
How would my stream name look like then?