I’m kind of sick of opening my tor browser everyday just for looking at important updates.
luckily if we get more people to join the fediverse, which RSS is an available option by default. in the future I might only need to just open my RSS feeds.
the problem is little to no content creators use the fediverse, even unaware that doing this is a good thing because of “First Mover Advantage”. on top of that, unlike that term doing this has no risks involved. you’re just expanding your audience.
People around the internet should ask their favorite youtubers and content creators to also use a Mastodon and Peertube account, maybe get them a crossposter software to do everything for them. if that exists.
like you and everyone else, I don’t want to rely on big social media corporations to connect with people.
I hope this message gets to somewhere, repost this or tell other’s to do the same.
to solve this problem we should take example from youtube.
read number 7 and 7.2; https://www.8bitmen.com/youtube-database-how-does-it-store-so-many-videos-without-running-out-of-storage-space/
but peertube is different from youtube, It’s federated. and we shouldn’t use a storage system like GFS/Bigtable, because it isn’t open source. instead we should use HDFS, and make a hosting datacenter designed for video streaming services. should be similar to this;
app = peertube client
this way, adding more storage would be easier and cheaper as you don’t have to move things around much. and instead of renting the drives they should have an alternative to buy an HDD storage tied to that account. a seperate ssd should be used for running the operating system.
The reason no hosting companies haven’t done this (or at least, I don’t know if it exists.) is because not much avid content creators host their own streaming site like peertube. this problem has never been a thing before.
The article is really interesting, thank you! But it’s all about performance it seems. As far as I can understand a decentralized network of small PeerTube instances don’t need much work for scaling, what we have to solve instead is rough storage size.
For sure we should improve the support for WebMonetization and get microdonations while streaming but again the main problem remain: the storage cost always increases over the time while the income is always tied to actual views/popularity/donations/whatever at a given time.
The video files have to be removed from the servers, the point is how. In addition to the archiving feature I described the “archived” videos could be streamed from the PCs of people making them available via (Web)Torrent. This should be techically possible since the support for WebTorrent is coming to libtorrent, the library used by many torrent desktop clients, but we would still need a lot of work on PeerTube side.
At that point the PeerTube instances would be a mere interface to stream video stored on people’s PCs and eventually caching popular videos automatically on the servers for better performance.
I think you missed my point here, yes It’s good for performance, but I’m talking about how a datacenter could run a video streaming service for multiple customers, if storage is the problem then they can put the videos in a HDFS (or GFS) file system. no matter where the storage is, they can assign it to a VPS user. similar to youtube datacenters work. doing this will be cheap and efficient.
that’s a great solution. even though it haven’t been implemented, at least they have something to work with.
Using a different file system can increase the performance but it doesn’t provide extra storage…
didn’t you read the article? it does make it 10x cheaper as it uses HDD only, and using that system we could add as much storage as we want to, if you don’t believe me this works, youtube uses this system. the only thing we need is a datacenter that is designed for that.
also take in the fact that if peertube finds a way to monetize videos, It may even become more profitable than youtube itself.
this will slow down that storage problem to the point that you wouldn’t even have to worry about it, because hard drives would eventually become alot cheaper.
edit; my mistake It’s not just a filesystem, It’s a data storage system.
The monetization can be made with the upcoming WebMonetization plugin.
The industry already provides storage on demand, see Amazon S3. I don’t think someone can build something better and convenient enough, the maintainance costs are huge too.
A YouTube channel with a few hundred subscribers may be able to make its PeerTube instance affordable if WebMonetization spreads enough and uses the PeerTube instance to deliver premium content and thus bypass YouTube’s membership system.
But providing the video upload service publicly to anyone like YouTube does is impossible because there is no such thing as Google’s ad system. Maybe in the future when the vast majority of users will have WebMonetization enabled.
I didn’t really payed attention to what you said there, you’re right, we’ll just have to get content creators to use peertube first.
For me the key is WebMonetization, if it will partially replace the advertising on the Web there are more chances that casual users will stream microdonations to PeerTube content creators automatically while watching their videos.
Coil’s implementation of WebMonetization already let its users support Twitch channels with microdonations. Imagine if at a certain point YouTube supports ad-free version of its videos if the user is streaming microdonations with WebMonetizarion. And most other premium Web services do the same, for example newspapers. At that point the WebMonetization userbase would be so huge that PeerTube instances can be supported by most of their visitors.
shshhh don’t let youtube know. they might realize that decision is a trap.