DDos (Distributed Denial of Service) attack against BlueSky a few days ago: The whole network goes down.
-
@agowa338 The TL;DR for me is that the current ATProto/Atmosphere architecture still contains centralised elements that stand in the way of full federation. My suspicion is that this is by design.
@jwildeboer@social.wildeboer.net @agowa338 Yes, the PLC Directory (Identitätsverwaltung) is (still) centralized.
-
@jwildeboer Oh they still are decentralized, but most users and the application/client aren't, so it seems moot.
It's worth remembering that Mastodon hit a per instance user limits early on, which led to a similar criticism that lasted years. Given enough time, today's network that doesn't solely rely on the official instance emerged.
It's also easy to forget that running a Mastodon instance can be EXPENSIVE! The one I'm on with maybe a few hundred active users costs $700/mo. IMO this should be way less, but I don't know if the Mastodon org can or has the right incentives to actually fix it.
There are HUGE benefits in both designs, but other than the outages the incentive for the BlueSky (AtProto) network to further decentralize aren't life-or-death as they were here.
@pawv The big difference to me is that running an activitypub based instance like mastodon (or gotosocial or misskey, peertube etc) means that everything is in one place. Sure, that might be several services/containers like the app itself, the database server etc. But it's one atomic unit. Participating in the ATProto/Atmosphere network means a whole collection of distributed services that can't be simply installed and run in an easy way, in some cases you need to rely on centralised parts.
-
DDos (Distributed Denial of Service) attack against BlueSky a few days ago: The whole network goes down.
DDoS against mastodon.social today: Lots of users on that instance impacted, the rest of the Fediverse keeps on running, my instance doesn't really notice, messages will be queued up until mastodon.social is reachable again.
Both networks claim to be decentralised and federated. You decide
@jwildeboer What's ultimately and unirrevocably unique that even during ddos attack I could use the basic UI features after certain time (less than an hour), and within ~2-3h posting worked with a bit of retries (usually less than 2 attempts). Fascinating!
I don't mind currently registering on bluesky, but their decentralization design is much weaker than this fediverse, despite having more public attention from influencers. -
DDos (Distributed Denial of Service) attack against BlueSky a few days ago: The whole network goes down.
DDoS against mastodon.social today: Lots of users on that instance impacted, the rest of the Fediverse keeps on running, my instance doesn't really notice, messages will be queued up until mastodon.social is reachable again.
Both networks claim to be decentralised and federated. You decide
@jwildeboer All I can say is dee doo dee doo dee doo. Decentralization is da bomb!
-
DDos (Distributed Denial of Service) attack against BlueSky a few days ago: The whole network goes down.
DDoS against mastodon.social today: Lots of users on that instance impacted, the rest of the Fediverse keeps on running, my instance doesn't really notice, messages will be queued up until mastodon.social is reachable again.
Both networks claim to be decentralised and federated. You decide
@jwildeboer There truly is a world of difference
I tried running a dedicated stand alone Bluesky server/network but i was unable to
It’s not as simple as just running a PDS, that’s just a super small part of it. Bluesky is designed to keep everyone needing their service one way or another
-
DDos (Distributed Denial of Service) attack against BlueSky a few days ago: The whole network goes down.
DDoS against mastodon.social today: Lots of users on that instance impacted, the rest of the Fediverse keeps on running, my instance doesn't really notice, messages will be queued up until mastodon.social is reachable again.
Both networks claim to be decentralised and federated. You decide
@jwildeboer was unaware of any such attack until reading this post lol
-
DDos (Distributed Denial of Service) attack against BlueSky a few days ago: The whole network goes down.
DDoS against mastodon.social today: Lots of users on that instance impacted, the rest of the Fediverse keeps on running, my instance doesn't really notice, messages will be queued up until mastodon.social is reachable again.
Both networks claim to be decentralised and federated. You decide
@jwildeboer ANy idea of where the attach came from?
-
DDos (Distributed Denial of Service) attack against BlueSky a few days ago: The whole network goes down.
DDoS against mastodon.social today: Lots of users on that instance impacted, the rest of the Fediverse keeps on running, my instance doesn't really notice, messages will be queued up until mastodon.social is reachable again.
Both networks claim to be decentralised and federated. You decide
@jwildeboer I had no idea dot social even got DDoS'd until just now.
-
@jwildeboer Who the heck would attack mastodon social??
@tech_noir @jwildeboer Many on the fedi actively hate mast.social (ban on sight kind of deal) I wouldn’t be surprised. It could also be a weird way to push people off .social to make mastodon more decentralized.
-
DDos (Distributed Denial of Service) attack against BlueSky a few days ago: The whole network goes down.
DDoS against mastodon.social today: Lots of users on that instance impacted, the rest of the Fediverse keeps on running, my instance doesn't really notice, messages will be queued up until mastodon.social is reachable again.
Both networks claim to be decentralised and federated. You decide
@jwildeboer
The whole network didn't go down. Bluesky went down, and non-bluesky architecture was fine. Let's acknowledge the other parts like blacksky, microcosm and brid.gy building the actual decentralization, and not conflate them with bs
-
DDos (Distributed Denial of Service) attack against BlueSky a few days ago: The whole network goes down.
DDoS against mastodon.social today: Lots of users on that instance impacted, the rest of the Fediverse keeps on running, my instance doesn't really notice, messages will be queued up until mastodon.social is reachable again.
Both networks claim to be decentralised and federated. You decide
@jwildeboer i never understand their logic about decentralize
-
And to be very clear about this: DDoS attacks are stupid and cause a lot of work. My sympathy goes to the sysadmins on every side, including those at the upstream network providers who have to deal with this, often at weird times. It is exhausting work and I really don't want you to have to do it, but hey, that's the world we live in, unfortunately. Sys Admin Hug!
@jwildeboer true. This shit is an unnecessary hassle for staff and disrupts daily routine. Ongoing projects have to wait and/or run into deadlines. And for what? Some kid that wants to cosplay as l337 h4xx0r
-
M monkee@other.li shared this topic
