I've had the idea of using a distributed IP blacklist using federated subscriptions ala diaspora and GNU social before and I think it might be the one solution

I've had the idea of using a distributed IP blacklist using federated subscriptions ala diaspora and GNU social before and I think it might be the one solution to actually have any kind of leverage in fighting abuse on the fail2ban-level for everyone who is not Google. There is probably still more work to do and this (IMO) does not scale to Layer 8 that easily.