Wowza Community

Redundant entry points @ Wowza Streaming Engine on EC2

I’m looking for the best way to set up primary and backup entry points for two encoders live streaming the same content, just like the way it is set up with Wowza Streaming Cloud or Youtube Live.

Our Wowza Streaming Engine server is running on Amazon EC2.

How do we best set this up and ensure an uninterrupted experience for those wathcing a live stream when one of the encoders fails?

Anyone knows how to set this up? I’m thinking of doing it one of these ways depending on the importance of the event:

Redundant encoder workflow

Two encoders streaming. Both to the same Wowza Streaming Engine server.

Redundant encoder and server workflow

Four encoders streaming. Two to server a. Two to server b.

To break the silence let’s focus on one thing then:

Encoder redundancy by having two live encoders streaming to the same Wowza Streaming Engine server, to the same application and to the same stream name:

Live encoder 1:

Server URL: rtmp://50.50.50.50/live

Stream name: [enc1]_myStream

Live encoder 2:

Server URL: rtmp://50.50.50.50/live

Stream name: [enc2]_myStream

Public streaming URL would then be:

http://50.50.50.50/live/myStream/playlist.m3u8

It would be prioritizing the primary incoming stream from encoder 1:

rtmp://50.50.50.50/live/[enc1]_myStream

If encoder 1 disconnects, the secondary stream takes over:

rtmp://50.50.50.50/live/[enc2]_myStream

Is it it possible to achieve encoder redundancy like this, without adding additional Wowza Streaming Engine servers into the equation?

My first idea would be to have two origin servers accepting streams from different encoders and have edge servers set up with primary and backup origin servers. But if you want to use only a single origin server, you could achieve the same thing using different vhosts. Maybe even simply using different applications in the same vhost, but I haven’t tried this.

However, if you’re trying to achieve redundancy for encoders themselves, I’m not sure this will work. If I remember well, origin failover works only if the origin server (application) itself is not operational, not sure if the failover would happen if only the stream is not available on the primary origin, but the origin itself is still operational (worth checking out, though).

In your last example, I’m not sure how the switchover would happen? If it happens on the web-application level, then it’s clear, but I’m not sure if there is any way to do it in Wowza itself.