Good question - I'm not sure what happens when the node storing the master playlist value goes offline. It's completely abstracted away from the user. One way we can account for this issue is by letting the broadcaster claim to be the "provider" of the data (I think this might be what IPFS uses for pinning), so that the broadcaster can serve as the fallback.
UpdateMasterPlaylist gets called every time there is a new variant for the stream. It gets called once per job instead of for every new segment.
Good question - I'm not sure what happens when the node storing the master playlist value goes offline. It's completely abstracted away from the user. One way we can account for this issue is by letting the broadcaster claim to be the "provider" of the data (I think this might be what IPFS uses for pinning), so that the broadcaster can serve as the fallback.
UpdateMasterPlaylist
gets called every time there is a new variant for the stream. It gets called once per job instead of for every new segment.