The Spam Problem.... new thoughts and ideas for control needed?

raza19

@joepie91 said:
cryptographical proof-of-work,

I got that but he is right, people would be selling computation power to do crypotgraphic hashing/puzzles then.

When I say captcha I mean captcha for validating the email as well as the sender once. After filling in captcha the system could communicate a private key valid for x number of days that will not require further captchas.

And for automated mails such a key could be provisioned to the genuine non-spamming mass mailing clients.

raza19

& btw its not like botnets have any shortage of processing power.

chihcherng

@joepie91 said:
EDIT: For anybody who doesn't feel like reading up on this; the mechanism is similar to that in Bitcoin, except here the sender of an e-mail completes the "proof of work" challenge, requiring CPU power, and making spamming less economically viable.

This may work if spammers send junk mail directly from their own computers, which is rare nowadays. Most spam are sent from botnets, thus the "proof of work" will be carried out by those malware-infected computers. Spammers probably won't notice any impact at all. This strategy might still slow down the transmission of both spam and legitimate mail.

joepie91

@raza19 said:
And for automated mails such a key could be provisioned to the genuine non-spamming mass mailing clients.

If I understand what you are suggesting, that creates just yet another single point of failure.

raza19 said: I got that but he is right, people would be selling computation power to do crypotgraphic hashing/puzzles then.

Doesn't matter. The goal isn't to make spamming impossible, but to make it not economically viable.

raza19 said: & btw its not like botnets have any shortage of processing power.

chihcherng said: This may work if spammers send junk mail directly from their own computers, which is rare nowadays. Most spam are sent from botnets, thus the "proof of work" will be carried out by those malware-infected computers. Spammers probably won't notice its impact at all. This strategy might still slow down the transmission of both spam and legitimate mail.

Not quite. If you look at the volume of spam that is actually sent from a single bot, that is going to require a lot of CPU power on the machines sending it. Likely enough CPU power to significantly slow down the machine it's sending from - which is effectively the fatal sin for a botnet operator, because detection means removal.

Bitcoin-mining botnets are rare for the same reason. It's hard to turn a noticeable profit from them at all. There are much easier ways (such as capturing bank data), that involve far less risk of detection.

DeftNerd

I've given up trying to send email from any VPS's. I just set them up to use a SMTP server and I use my free (for reasonable usage levels) mailgun.com account.