User will upload the videos from browser which need to be stored on server and playback. First understanding from google suggest I need to go for HTTP live streaming(HLS) here.
But I am not sure how it works internally ?
There are three components in above workflow i.e. client/server/data store for saving and retrieving videos.
Save flow :
I believe I need to plugin the HLS client for sending the streaming data.
Does Client in itself divide the file into chunks while sending and maintain the chaining of these chunk where each chunks points to next one ? something like this as I believe server is dumb and will work in same fashion as http upload functionality and no other intelligence is required here ?
But not sure how HLS server side component works here i.e. Will it save as single file or single file is split in to multiple files and then saved on disk ?
I believe it store the file as single file like regular http upload file ?
Retrieval part
In normal regular http file download, client asks for file data and server sends the response back in chunks but all response chunks are sent back against the same request.
I believe in case of HLS , its pull based where client initiate the pull request for each stream request. In each chunk pull request client gets the file name of next chunk and send the request to serverthe relevant chunk from single file for each poll request etc ? So for server its kind of regular http file download request and all intelligence lies with client
与恶龙缠斗过久,自身亦成为恶龙；凝视深渊过久,深渊将回以凝视…
