Are the registry procedures causing the problem? Since most of my communities are on lemmy.ml, wouldn’t I still be using lemmy.lm’s bandwidth? Or is it more crucial to ensure that the communities are distributed more evenly across different instances?
To get serious about distribution, I think there needs to be a few features implemented;
- User migration. Allow people who find themselves on an overstuffed server to migrate to a different instance
- Community grouping / merging. If there are multiple communities for the exact same thing, allow the mods to group them or merge them to not fracture the user base
- User data import / export or cloud backup to a personal provider. If one day your server instance just disappears, it would be nice to know I have a backup of my user data, subscribed communities, saved posts and comments, and maybe a way to relink to my old posts and comments across the fediverse.
Just a few ideas I was thinking about that could possibly help users move around.
I find it strange that anything lives on a single server at all. It’s wasting one of the best parts of being peer to peer. Redundancy. Every server should just be sharing the ultimate load and hosting parity so nothing can be truly lost when someone decides they’re done with their side project…
Ah, I misunderstood then.
What happens if lemmy.ml is down?
All the users and communities registered there will cease being able to connect to the rest of the network, but the data should still be cached on all the other servers.
The keyword is “cache”. Lemmy doesn’t guarantee the cache, so if an instance X would go down permanently, you would eventually lose this data.
My presumption was that with over 200 instances, maybe over 1000 in the near future, all the old data should be collectable in theory without loss. But you’re right, there’s no guarantee of this.
That’s what federation means, not really peer to peer
Those are exactly my thoughts. Migration and merging communities will be important sooner or later
I would add to this community migration, which will be important as instances start going offline. User migration is great, but, whereas on Mastodon, the content lives on the user, I believe here it lives on the community.
I think we want communities to spread out more than users.
Unfortunately if the users are there then they create the communities. So really people just need to spread out.
Migrating means losing a few things you’ve made or commented on. We’re so young that shouldn’t be such a big deal yet though.
This l load balancing will have to be automated sooner or later (sooner hopefully) because it’s gonna be problematic…
Yeah. I like being able to join an instance. But I would really like to be able to backup my profile and to be able to transfer it in the future. It would be ideal to be able to just spin up a raspberry pi, and transfer my account to it at some point.
I think we want communities to spread out more than users.
Unfortunately if the users are there then they create the communities. So really people just need to spread out.
Migrating means losing a few things you’ve made or commented on. We’re so young that shouldn’t be such a big deal yet though.
This l load balancing will have to be automated sooner or later (sooner hopefully) because it’s gonna be problematic…
Did you intend to duplicate your post so many times or is it just the server messing things up?
Server and app weirdness. I deleted the duplicates. But I guess they have to prepare out to the other servers.
That’s a good idea! but which instances would be appropriate?
Literally any. Browse through the instance list and find one that matches whatever criteria you think is important to you.
Im French Canadian and when I saw that sh.itjust.works was a French Canadian instance I figured why not?
It’s not super important stove you can see everything in the network from all instances unless it has specific rules against something.
Just my thoughts from a random user
-
Have communities work similar to how RAID works with drives. It spans across multiple trusted instances, with data mirrored or striped to each. If one goes down, a new instance could become trusted and have the data rebuild on that instance. This would create redundancy and prevent one instance from becoming the main or default.
-
Allow similar with user accounts or allow instance migration.
-
Yeah, the faster Lemmy will figure out an account migration the better, i applied for Lemmy.ml account before I’ve noticed the overload post, now I’m stuck with an intense that’s a bit more laggy than I like
I think we want communities to spread out more than users.
Unfortunately if the users are there then they create the communities. So really people just need to spread out.
Migrating means losing a few things you’ve made or commented on. We’re so young that shouldn’t be such a big deal yet though.
This l load balancing will have to be automated sooner or later (sooner hopefully) because it’s gonna be problematic…
