The server itself is great, splitting services out is FAB. The content mirroring… chefs kiss.
To be entirely clear, what’s missing and what we need is a Mac, Windows and Linux stand alone app. (low configuration) You point it to your videos from local storage/network and they become a locally hosted resource (torrent like), You then connect to a real hosted server, where your indexed, media and meta are populated. Your app gets port forwarded, so the public server is just indexing you and pointing people to you.
If something happens to the public server you’re on, you point to a new public server and all your content still exists. Pirate radio style.
Since you log into the public server as a user, they can still moderate you as they see fit, block your content, or mirror your content to their server with the already existing features. If you don’t like how they handle you, you can move to another server, or host your own somewhere.
This puts the onus of base storage and the first hop of network connectivity on the content maker. But disks are cheap, and the network is peer-based, so if you get popular, your own watchers will help each other out.
Then you might as well just host your own PeerTube instance, which is also something I would encourage big content creators to do. Problem is, you would have to be a bit tech savvy to install and run PeerTube.
The standalone app you talk about probably doesn’t work in practice. What happens when the user shuts down their computer? The videos become unavailable. Then what?
What could work is a 2-in-1 solution, where a content creators video backup also functions as a peertube instance.
Content creators back up their videos, so why not make the videos available for watching, from the same storage?
Then you might as well just host your own PeerTube instance
Your average person is not capable of setting up, hosting, backing up and maintaining a proper Peertube instance. A large part of my dayjob is in hosting and infra, and I have set up a PT instance before. Honestly, even as far as average hosting goes, their setup is kind of needy. Most dockers are like start, port, go these days. As soon as you mandate that they have DNS and a working 443 upfront, you’re kicking most people out before they start.
The standalone app you talk about probably doesn’t work in practice.
There is no challenge there that cannot be overcome by application architecture. I don’t see anything that is a deal braker, and there is definitely a need out there, but it’s a damn big project and the makers of PT aren’t just doing straight up charity.
What happens when the user shuts down their computer?
Indexer loses heartbeat, content shows up as unavailable. Configurable by admin to disappear from searches after x time. Current watchers: If enough people were watching the stream currently/recently, they’d keep going from each other’s cache as peertube will actually peer. (I’ve tested the peer part, it does what’s on the tin) If no one has future segments, they’d get kicked out same as if the server went offline. Most content creators would be pretty careful not to do this. It would also be an interesting thing to explore the mirroring functionality and have small creators team up and mirror each others data, doing a kind of remote replica between friends thing.
What could work is a 2-in-1 solution, where a content creators video backup also functions as a peertube instance.
That’s actually kind of my point. It’s mostly Peertube, but without users trying to deal with nginx/apache, redis/postgres, port 80/443 ips issues, SMTP, DNS and SSL. Someone who is capable of dealing with those parts just stitches them on as remote content. The servers don’t pay for storage or most of the networking and they help the clients with visibility and getting around ISP limits.
If we try to setup end users with the whole shebang, they’re going to run into issues when the Redis goes RO from running out of space/memory, or when they have disk issues and the postgres goes tits up. (or god forbid PT needs a DB upgrade or something) They don’t know about services and recovery, they’re not capable of setting up backups or running restores.
Hell maybe the app is also the consumption part as well. You watch peertube by running the PT desktop app. Then we can deal with self signed certs. run people on sqlite and local shared host memory instead of redis. We’d still need remote trackers to help orchestrate the DHT.
I duuno, it’s big, it’s pain. I think if it were available, big tech (and honestly big brother) would be shitting themselves.
The server itself is great, splitting services out is FAB. The content mirroring… chefs kiss.
To be entirely clear, what’s missing and what we need is a Mac, Windows and Linux stand alone app. (low configuration) You point it to your videos from local storage/network and they become a locally hosted resource (torrent like), You then connect to a real hosted server, where your indexed, media and meta are populated. Your app gets port forwarded, so the public server is just indexing you and pointing people to you.
If something happens to the public server you’re on, you point to a new public server and all your content still exists. Pirate radio style.
Since you log into the public server as a user, they can still moderate you as they see fit, block your content, or mirror your content to their server with the already existing features. If you don’t like how they handle you, you can move to another server, or host your own somewhere.
This puts the onus of base storage and the first hop of network connectivity on the content maker. But disks are cheap, and the network is peer-based, so if you get popular, your own watchers will help each other out.
Then you might as well just host your own PeerTube instance, which is also something I would encourage big content creators to do. Problem is, you would have to be a bit tech savvy to install and run PeerTube.
The standalone app you talk about probably doesn’t work in practice. What happens when the user shuts down their computer? The videos become unavailable. Then what?
What could work is a 2-in-1 solution, where a content creators video backup also functions as a peertube instance.
Content creators back up their videos, so why not make the videos available for watching, from the same storage?
Your average person is not capable of setting up, hosting, backing up and maintaining a proper Peertube instance. A large part of my dayjob is in hosting and infra, and I have set up a PT instance before. Honestly, even as far as average hosting goes, their setup is kind of needy. Most dockers are like start, port, go these days. As soon as you mandate that they have DNS and a working 443 upfront, you’re kicking most people out before they start.
There is no challenge there that cannot be overcome by application architecture. I don’t see anything that is a deal braker, and there is definitely a need out there, but it’s a damn big project and the makers of PT aren’t just doing straight up charity.
Indexer loses heartbeat, content shows up as unavailable. Configurable by admin to disappear from searches after x time. Current watchers: If enough people were watching the stream currently/recently, they’d keep going from each other’s cache as peertube will actually peer. (I’ve tested the peer part, it does what’s on the tin) If no one has future segments, they’d get kicked out same as if the server went offline. Most content creators would be pretty careful not to do this. It would also be an interesting thing to explore the mirroring functionality and have small creators team up and mirror each others data, doing a kind of remote replica between friends thing.
That’s actually kind of my point. It’s mostly Peertube, but without users trying to deal with nginx/apache, redis/postgres, port 80/443 ips issues, SMTP, DNS and SSL. Someone who is capable of dealing with those parts just stitches them on as remote content. The servers don’t pay for storage or most of the networking and they help the clients with visibility and getting around ISP limits.
If we try to setup end users with the whole shebang, they’re going to run into issues when the Redis goes RO from running out of space/memory, or when they have disk issues and the postgres goes tits up. (or god forbid PT needs a DB upgrade or something) They don’t know about services and recovery, they’re not capable of setting up backups or running restores.
Hell maybe the app is also the consumption part as well. You watch peertube by running the PT desktop app. Then we can deal with self signed certs. run people on sqlite and local shared host memory instead of redis. We’d still need remote trackers to help orchestrate the DHT.
I duuno, it’s big, it’s pain. I think if it were available, big tech (and honestly big brother) would be shitting themselves.