An IBM scientist has an old new idea as reported on Wired News: Charge spammers a nuisance fee.

The net is basically free today, and making people pay for anything has been an uphill struggle. The most successful campaign has been the RIAA/MPAA campaign against napster et al, and even that is a partial failure.

This anti-spam concept is misguided for two reasons. First of all: open email - even with the spam - is still a benefit to the consumer, and spam is simply not a big enough problem to make 'priority email' a killer-app. If you don't think so, you simply have forgotten what the world was like without email.

If this assumption was wrong in any significant way, mail clients with listed senders would be out there. It is entirely possible to stop spam completely using an intelligent mail client verifying senders, and issuing an 'I don't know you, so I won't receive your email' receipt to unknown senders automatically. Nobody I know uses such a system.

But even if my assumption was wrong, a scheme as radical as the one proposed without strong corporate sponsorship is unlikely to succeed, and corporate sponsorship of traffic limiting technology is hard to see as a very possible event. It could only work as an ISP value add or something like that, and it seems to me that the main theme of the internet today is that nobody cares who their ISP is these days - hence AOL's current problems.
We're left with the slight possibility that the OS monopoly power would get into this as a new way of extracting revenue from users.