Alright so I’m not an expert so I might not be explaining it correctly.
Federated Network: Multiple instances sharing content, such as Lemmy
Peer to Peer Network: There is no “instances”, just peers. Many peers sharing content. Every user is a peer. There is no server costs, because every device connected to the network is acting like a mini-server. It will cost your device some storage space and network bandwith depending on the how the software is designed.
Or do you think Centralized servers are still gonna dominate the future?
With distributed hash tables it is manageable. You do something like “store three copies on three peers” and as long as one of them is online the post is accessible. This is actually better than the way lemmy does it now. In principle each lemmy server stores the posts from its communities, and a copy of each post from communities its users are subscribed to. But since all instances are federated so well, in practice each of the 1000 lemmy instances stores a copy of almost every post ever made. That’s like 100GB x1000. With a DHT, the amount of space used on each user’s device is on average the amount of posts one user makes x3, no more.
Yeah I have serious concerns about how it will scale.
Luckily storage is the cheapest thing generally.
Maybe down the line they can start using varying degrees of cold storage for older content. Cheaper to store but more expansive to access.
Storage is cheap but most of the instance operators that are setting up right now aren’t prepared for how much storage they’re going to need and it’s associate costs. I’m not talking the big boys like .World, but the hundreds of private and semi-private instances being set up on $12/mo VPS and such.
After 30 days of running my single user instance I’m at 23GB of storage. Since I’m using on prem equipment I have the lowest cost per GB possible and am not the least concerned. We’re going to see a ton of attrition with hosted instances as the costs of ownership goes from 10, to 20, to 40, to $50+/mo. due to storage. Many aren’t in anyway prepared to tackle the topic of moving PICTRS off to object storage or engage in other mitigations.
I’m not worried. Text is tiny to store and any image posted remains on the source instance and is not duplicated, just linked.
We could also just delete stuff after some time. Nobody really needs the 1000th repost of a meme from 20 years ago.
If they’re storing it right, it is content addressed, meaning that their servers are aware of duplicates because each file is hashed.