Be that as it may, there must be something we can do about this as relay operators. If you get caught doing something illegal on your home Internet connection, there are warnings, and eventually consequences (like being disconnected). Just because you run a Tor relay doesn't mean the rules don't apply to you, and if we can't do anything to stop illegal activity, eventually relays are going to be disconnected.
I understand both sides of the argument, and why no solution would be perfect, but we need to figure something out. This problem will not go away on its own, and I expect it to only get worse as time goes on.
Personally, I don't like the idea of filtering traffic at the exit node, because it seems to undermine the whole purpose of Tor: unrestricted anonymous access. However, there must be some way to identify at least some malicious traffic, such as bots. If Tor relays start filtering traffic, I think it should be opt-in, and it should happen at the guard relay. That way not all relays filter by default, and if something gets blocked, it happens
before it gets routed through the network.
Of course, we could always identify what constitutes as filtering. As already stated, each exit relay has its own exit policy, so technically everyone already filters traffic based on port. If an IPS only logs non-identifiable information, I don't think it would compromise anonymity, but at the same time, people may not trust Tor if it starts scanning traffic.