I remember we discussed something alike a few years ago, but it wasn't really stable as it was mostly depending on how many viewers you had and each viewer was a peer, so if you had little viewers or the viewers where too far away from each other you would never have a stable connection
I now mostly stream on Twitch for another project I'm working on. But I gladly shift to VIMM on Hive if it helps. The idea sounds fascinating, is it a chain on it's own (and mineable) or is hashpower aquired using renting services like Nicehash? Also how many streams a week on average should testers aim for?
I remember we tested a P2P delivery system a while ago, but we didn't notice any improvement. Heck, some streams were actually worse when there was only a few viewers.
But this is a completely different approach. The delivery is achieved using a standard CDN for smooth playback. The part that is really decentralized here is all the GPU power used for transcoding. It is built on the Ethereum blockchain. The "miners" are basically allowing the network to use their GPU power to transcode streams, and in return they get ETH and LPT (Livepeer Tokens).
This allows us to scale at a fraction of the cost of setting up traditional ingest servers, while delivering better performance. There's also a few cool things like the clever load balancing that automatically switches you to a different ingest if the one you're connected to is performing poorly.
Honestly this is just like playing with a shiny new toy.
This sounds interesting, this is the first decentralized solution for video feed that sounds like it works (though I have to admit I don't read/research a lot into it). If it is as interesting I might even look into "mining" it as well
Sign me up
We're just looking to get as many people from different regions on the new ingests for now, the weekly average doesn't matter as much.