Re: Macro payments through domain names?
Marc Mengel <[email protected]>
| Newsgroups | gmane.mail.im2000 |
|---|---|
| Organization | Fermilab Core Servers and Infrastructure Group |
| Message-ID | <[email protected]> |
[email protected] wrote: > Put the hash (eg., md5) of a public pgp/gpg key used for signing > emails in a DNS entry derivable from the email address. Enforce the > DNS lookup to be in two steps. The first lookup returns the reference > to an external top level domain; the 'eMail Authorization Domain' > (mad). The second lookup is a TXT query towards this 'mad' which must > return the hash of the public key of the email signature (where the > public key and the signature data is included in the end of the email > as usual). > > Now use a system of blacklists similar to how it works today, but for > sent spam, blacklist the involved 'mad' instead of the domain of the > sending email address. This makes a 'mad' used for spamming > "unusable" for further emailing, and someone will have to pay for a > new domain that isn't yet on the blacklists; but when doing so one > will be able to keep using the same email address as before (which is > important if some other user under your domain was responsible for > the spamming). I was thinking of something only a little different; you get to buy (rent ?) a key that's good for N emails/month. This does mean that a counting service would be needed, which would count reported email receipts for each key. Exceeding the purchased rate *automatically* blacklists the key for the rest of the hour/day/month -- so it provides throttling against even unintentional excessive emailing. That way, anyone who wants to run their own mail service can buy (rent?) a key, and ISP's could build the cost of key rental for their clients into the service agreement price. To prevent false reporting of delivered email (which could otherwise allow a denial of service attack) the sender would need to send in a (message-id, recipient) list to the counting service (signed). The recipient report would then be checked against the pending message-id,recipient pair, and the mail count would be updated. > Spammers paying for their own 'mad's will have a hard time to make > profit, and will be traceable by whois data. Spammers using hacked > systems will really push people towards higher computer security > since such incidents now leads to a real monetary cost. However, the > actual cost will hardly drive anyone out of business, since it can > be limited to the registration fee of one domain per organization. If your system is hacked; you could pay a key-replacement fee, and get a new key issued.