Recently I'm doing some work on RTMP streaming, that is using Flowplayer to integrate with Edgecast Streaming service and CloudFront Streaming service.
The basic concept is easy to follow, but the format of different providers really waste me a lot of time to figure out.
For example, in order to make edgecast happy, according to the documentation, you need to specify filename in the format of mp4:filename.mp4, flv:filename (without .flv extension) and mp3:filename (without .mp3 extension).
But for CloudFront, it's a different story that mp4:filename.mp4, filename (no flv:prefix, and no .flv extension) and mp3:filename (without .mp3 extension).
This format makes people even more frustrating when today I try to use Edgecast's loadToEdge function, the format the accept is filename.mp4 (without mp4: prefix), filename.flv (without flv: prefix) and mp3:filename.mp3.
As you can see, basically there is no logic there and you have to guess and try all different combinations to make it finally working.
I just would like to know if anyone has idea about why different providers implement their streaming in all customized way? Or is it Adobe's fault doesn't have a unified form or it's just up to service providers to use whatever they like.
Thank you!
It's all about implementation. The URL format, including extensions, has nothing to do with
As an analogy, your question is like asking "Why do some websites have different URLs than others?" Example of two different yet viable ways of serving up an image:
http://server.com/question/87/why/65.pnghttp://server.com/image/question?number=87&image=65. It's all about how the coders at EdgeCast, Amazon, et al wanted to implement their CDN. I'm sure there was some logic to it, well thought out or not. And probably some need to deal with legacy systems, clients and URLs.
It has nothing to do with FMS itself. Just like the above analogy's URLs have nothing to do with the web server they are served from.