Some people love saying Mastodon/OStatus is very scalable because of federation but let's be honest that's some bullshit. It's easy to split and multiply, but that's not scaling.
The birdsite is much easier to scale.
Just imagine one edge case we're not seeing for now: many people using it (let's say millions) and following a very famous person
That person must have their own instance at this point, at least.
... what happens when they toot?
just imagine the enormous peaks and lag it'll do. And we're not even talking about media.
How would Mastodon realistically handle Katy Perry? I have no idea.
Sure it will work without that edge case, it's not *necessary* to have very famous people, but current society and all that.
that's something really important for the success of a social network these days. If something's gonna "kill" (=slow its evolution, really) it's probably that
@djsundog yes, that'd be very interresting
@CobaltVelvet because if the network /can/ handle it, having a high profile account with personal/fan base instance with high profile account support staff moderating the instance could be highly attractive to people who now have to entrust maintaining those connections via a third party provider like twtr.
@CobaltVelvet I think the on-boarding process might be a hitch, too. The main instance needs to be available enough that Katy Perry can just wander by and set it up on a whim and start tooting; right now there's a bunch of confusing decisions to make right off the bat. The penalty for getting it wrong is relatively low, but for some people, they won't stick it out through the first migration.
Or maybe Katy's IT guy will make a bespoke instance for her and that'll be that.
@CobaltVelvet this is something I'd really like to simulate - set up a test fediverse of N instances on VMs on a closed network, populate them with userbases, rig the follow graph of one, set the others to boost/fav/expand/etc. and see what the numbers look like during propagation like you describe.