Here's a few things you can expect when there's a #Fediverse outage
-
Here's a few things you can expect when there's a #Fediverse outage:
- oh wait, there's no such thing, the Fediverse is actually federated
- there are over 30k totally independent servers
- if one of them goes down the rest of the Fediverse works just fineRE: https://bsky.app/profile/did:plc:ragtjsm2j2vknwkz3zp4oxrd/post/3lawdqryaws2q
-
@caesar I understand your point, but... if there's an internet backbone issue and you can't reach your chosen server, then that's a Fediverse outage for *you*. If there's a fiber cut in Germany (where mastodon.social and other instances are hosted), a significant percentage of the Fediverse will be inaccessible to users during that outage.
-
@steve If I'm reading you correctly you're saying that the Fediverse has single points of failure because the internet does?
Whilst I can't argue with that, it doesn't really point to a flaw in the federation model of the Fediverse. As compared to the centralised model of Bluesky (both the canonical identity authority and the relay system are totally centralised), the federation model of the Fediverse is much more resilient to outages affecting the whole network. -
@caesar No, my point was that, as an individual user, the effect is the same (even if some unrelated part of the AP network is still operating in some manner).
-
@steve I guess if you want to avoid that kind of failure mode altogether, you need a fully or mostly P2P model like SSB or Nostr. But every system has its downsides… My original post was simply a light-hearted parody of Paul Frazee's post, intended to highlight that the Fediverse isn't susceptible to outages as a whole in the way that Bluesky is. Of course individual nodes can go down, but the rest of the network isn't affected, it will never all go offline at once.
