<$BlogRSDUrl$>

Friday, December 26, 2003

Sending Party Pays 

Who hasn't been pained by spam, especially if one happens to own a web-based e-mail account? Researchers have been trying to figure ways to fight spam for some time now. These have included several attempts to make spammers pay an economic cost for the e-mails they send out. The BBC is carrying a story about Microsoft's attempt to make sending parties pay for e-mail.

The development has been called the Penny Black project, because it works on the idea that revolutionised the British postage system in the 1830s - that senders of mail should have to pay for it, not whoever is on the receiving end.

"The basic idea is that we are trying to shift the equation to make it possible and necessary for a sender to 'pay' for e-mail," explained Ted Wobber of the Microsoft Research group (MSR). The payment is not made in the currency of money, but in the memory and the computer power required to work out cryptographic puzzles. "For any piece of e-mail I send, it will take a small amount computing power of about 10 to 20 seconds." "If I don't know you, I have to prove to you that I have spent a little bit of time in resources to send you that e-mail. "When you see that proof, you treat that message with more priority."

Mr Wobber and his group calculated that if there are 80,000 seconds in a day, a computational "price" of a 10-second levy would mean spammers would only be able to send about 8,000 messages a day, at most. "Spammers are sending tens of millions of e-mails, so if they had to do that with all the messages, they would have to invest heavily in machines." As a result of this extra investment, spamming would become less profitable because costs would skyrocket in order to send as many e-mails.

The idea was originally formulated to use CPU memory cycles by team member Cynthia Dwork in 1992. But they soon realised it was better to use memory latency - the time it takes for the computer's processor to get information from its memory chip - than CPU power. That way, it does not matter how old or new a computer is because the system does not rely on processor chip speeds, which can improve at rapid rates. A cryptographic puzzle that is simple enough not to bog down the processor too much, but that requires information to be accessed from memory, levels the difference between older and newer computers.


Seems like a clever idea, if it only weren't Microsoft promoting the idea. Anyone who has used Hotmail knows how much the volume of spam increased since the time Sabeer Bhatia sold to MS. Besides, what to do about the open standards that might be required to make even the Penny Black project work efficiently?