But the new users can join any lemmy server and still interact with everyone, so ideally the load would be distributed throughout all the instances instead of making one of them overloaded.
Yes but even 40k users per instance requires substantial amount of resources right now to keep things not-stuttering. And Apollo has millions of users. So we’d need 200 new instances as big as the largest ones we already have. And that’s assuming activity pub caching is already ironed out. Idk if it is.
That’s a good point! I think we usually hear about millions of users and don’t realize the scale you need to serve the thousands that’d be in each instance…
But the new users can join any lemmy server and still interact with everyone, so ideally the load would be distributed throughout all the instances instead of making one of them overloaded.
Yes but even 40k users per instance requires substantial amount of resources right now to keep things not-stuttering. And Apollo has millions of users. So we’d need 200 new instances as big as the largest ones we already have. And that’s assuming activity pub caching is already ironed out. Idk if it is.
That’s a good point! I think we usually hear about millions of users and don’t realize the scale you need to serve the thousands that’d be in each instance…