@volkris @lps @errhead @Revertron @kzimmermann What you're saying is not true. Smaller instances don't consume "exponentially more resources". ActivityPub is very easy on resources and servers can run almost anywhere (unless you use some really inefficient bloatware)
Conversation
Notices
-
silverpill (silverpill@mitra.social)'s status on Tuesday, 21-Nov-2023 04:31:28 JST 
silverpill
-
volkris@qoto.org's status on Tuesday, 21-Nov-2023 07:42:40 JST 
volkris
@silverpill if you read through the ActivityPub standard, the protocol is designed to scale with number of instances, and it is a scaling factor larger than simply linear.
You say AP is very easy on resources, but that flies in the face of so many instance operators experiencing meltdowns and surprise hosting costs when the resources exploded far higher than they expected.
Which makes sense because, again, the protocol is designed in a way that involves exponential increases.
-
silverpill (silverpill@mitra.social)'s status on Tuesday, 21-Nov-2023 07:43:09 JST
silverpill
@volkris @lps @errhead @Revertron @kzimmermann
Yes, AP is very easy on resources. I've built a full-featured AP server from scratch so I know how the protocol works and how it behaves under different conditions.
In AP world you only receive stuff you care about, so most of the time server is just waiting for something to land into inbox, or for outgoing request to complete. There's nothing else to do
>so many instance operators experiencing meltdowns and surprise hosting costs when the resources exploded far higher than they expected
Tell them to use better software.
-
All ０７６萌ＳＮＳ content and data are available under the