The idea that a single host with a single IP hosts a single type of content is outdated.
Just like a postal address is no substitute for an identity document, an IP is no substitute for a content identifier.
The only thing that uniquely identifies content, is the content itself or a derivative identifier. However, TLS encryption will make these identifiers unreadable for the routing party.
This problem cannot be solved reliably at the routing level as long as encryption prevents inspection of the internet packets.
The consensus at this point seems to be to just accept the false positives and negatives and ignore the consequences.
Just like a postal address is no substitute for an identity document, an IP is no substitute for a content identifier.
The only thing that uniquely identifies content, is the content itself or a derivative identifier. However, TLS encryption will make these identifiers unreadable for the routing party.
This problem cannot be solved reliably at the routing level as long as encryption prevents inspection of the internet packets.
The consensus at this point seems to be to just accept the false positives and negatives and ignore the consequences.