With Nginx RTMP there's an interesting feature that it's dynamic pull, with that an stream is served from a remote origin only when an user are requesting this stream but it only works with if an user request an rtmp stream, instead if the first request is HLS it doesn't work (it generates the HLS fragments after the rtmp stream is on play). I need to know if Wowza could work in that way but allowing this kind of workflow with HLS. Thanks in advance.
Using Wowza Streaming Engine in an Origin/Edge repeater workflow, this "
On Demand Mode" is the default configuration. When the first request for a stream is received on the Edge server, a pull request is made to the origin and the stream is initialized. If the first request is for an HTTP based stream, then there is a short delay while the packetizing initializes. This only occurs on the first request and can be reduced by
adjusting the packetizer settings.