To me, this feedback is the invaluable missing link that
all the third-party trust and shared-trust schemes I've
seen lack. When fred maintains his own trust keyring,
both fred and the bank know where the spammer got the
trust token. Also, both of them know the other will know
where the spammer got the trust token. A spammer getting
the trust token fred gave the bank will cost the bank
money, customer goodwill, and potentially business. And
fred and the bank must have contact with each other about
it if there is still a mutual wish to reestablish email
communication.
Machines suborned by spammers would be detected instantly
because all the trust tokens would be revoked within minutes
of the spammer using them. Spammers could not share trust
tokens around to get thousands of spam delivered on each
one, because fred would revoke the token the first time he
got spam on it and all the other spam would disappear. This
would destroy spammers' motive to cooperate by selling each
other spamlists. And when fred did revoke a key, none of
his legitimate email still arriving under non-compromised
tokens would be affected. Literally *EVERY* piece of spam
would be evidence of an information security breach traceable
to a single entity that fred gave his address to, and every
time spam were recieved there would be a single action that
fred could take to stop getting spam derived from that
particular information security breach.
This is basic digital signatures; it would work. We know
how to do it. We've known how to do it for a long time.
People would need to add one field to their email information
to keep trust tokens. Everything else you could implement
in the Mail User Agents. My inner cynic says nobody's
implemented it because nobody can figure a way to make money
by selling trust when users manage their own keyrings rather
than paying a third party to manage keyrings for them. Can
anybody else think of a better reason, or are we really
just lazy greedy bums who can't be bothered to work on
something unless we can "make money fast" doing it?
Bear
https://www.metzdowd.com/pipermail/cryptography/2009-January.txt
all the third-party trust and shared-trust schemes I've
seen lack. When fred maintains his own trust keyring,
both fred and the bank know where the spammer got the
trust token. Also, both of them know the other will know
where the spammer got the trust token. A spammer getting
the trust token fred gave the bank will cost the bank
money, customer goodwill, and potentially business. And
fred and the bank must have contact with each other about
it if there is still a mutual wish to reestablish email
communication.
Machines suborned by spammers would be detected instantly
because all the trust tokens would be revoked within minutes
of the spammer using them. Spammers could not share trust
tokens around to get thousands of spam delivered on each
one, because fred would revoke the token the first time he
got spam on it and all the other spam would disappear. This
would destroy spammers' motive to cooperate by selling each
other spamlists. And when fred did revoke a key, none of
his legitimate email still arriving under non-compromised
tokens would be affected. Literally *EVERY* piece of spam
would be evidence of an information security breach traceable
to a single entity that fred gave his address to, and every
time spam were recieved there would be a single action that
fred could take to stop getting spam derived from that
particular information security breach.
This is basic digital signatures; it would work. We know
how to do it. We've known how to do it for a long time.
People would need to add one field to their email information
to keep trust tokens. Everything else you could implement
in the Mail User Agents. My inner cynic says nobody's
implemented it because nobody can figure a way to make money
by selling trust when users manage their own keyrings rather
than paying a third party to manage keyrings for them. Can
anybody else think of a better reason, or are we really
just lazy greedy bums who can't be bothered to work on
something unless we can "make money fast" doing it?
Bear
https://www.metzdowd.com/pipermail/cryptography/2009-January.txt
Comments
Be the first to add a comment
Log in to comment