I’ve occasionally wondered why the responsibility to mitigate abuse can’t be pushed all the way out to edge nodes (individual ISPs and data centers) instead of relying on centralized abuse prevention. I imagine this does happen for certain kinds of abuse, but wouldn’t the threat of others not peering with your node be good motivation for ISPs not to provide services to spammers?
That's one big part of what Google does, and it's part of why it's so hard to get Google to accept self hosted email. They have a list of known good hosts (that they effectively "peer" with), and if you're not on that list your mail may or may not make it into a Gmail inbox.
