1432
Welp that answers a lot of why all .ml are down
(i.imgur.com)
This is a most excellent place for technology news and articles.
Man, hacking, DDOS and now this. The fediverse just can't catch a break...
Would help if users spread out over all the running servers because problem is just a few lemmy servers have all the users. For example the instance I run would be a simple proxy to use for all the content and then would mitigate issues when a big server had problems since just parts of the fediverse would be affected from the users pov.
Does that really scale though? The load on a server is not dependent on the number of users, but on the number of communities from other server that the sum of user is subscribing to.
Which means if you have a server for 100 users, you still need to pay for the 1000s giant communities that those users are subscribing to, as they are being copied over in your server.
So if you have a few mega server like Lemmy.world, they each pay say 10000£ in hosting a month (number taken out of my hat), which is fine because they have as many users that can contribute to it financially ( via donations, ads etc.). But small servers won't be able to support that load and will ultimately close.
That sounds like a design flaw if you ask me but i did not see anyone mentioning it so maybe i'm misunderstanding.
No its not really as bad as that at all. The disk space is linear in that way but disk space is cheap. All the rest is not taxed heavily by federation. Do the big costs like CPU dont scale up like that.