If there is an issue with the DSA licensing scheme, I believe this method
could use any digital signing method in order to get the desired effect.
Agreed. The RSA patent has expired, so I would guess it is a better
solution (all other aspects of this solution ignored).
The difficulty alone in solving the computation task makes it impossible
for a single CPU or group of them to be effective at spamming. If mail
host administrators generally picked a 10-20 second work load for a message
to be sent, that would mean 10-20 seconds between each message which would
be a huge detrement to spamming.
Yes, I understand the concept. However, I would imagine that even if
every single email person in the world adopted this approach, there
needs to be a way for users who know each other to distrubute keys to
each other. That would certainly be the case for a person signing up
to email for the very first time. The obvious way to distribute this
is to send an email with their PKI public key attached, and to do
this, they would need to resort to the challenge response protocol. So
the challenge response protocol could never be too constraining,
because people's first experience of email would be attrocious. Given
that, I think that spammers who thought they had a very valuable
product may still choose wear the CPU costs of going through the
challenge response protocol.
Well, that is a good idea however I think this "problem" would be solved
with a good implementation. If the mail host generated a random message
and required the client to sign it, there would be no way for a client to
know what information it would be required to sign and no way for other
clients to use them.
I'm a litte confused by what you are saying here. From my reading of
http://meor.xwarzone.com/overview.htm, it doesn't mention anything at
all about the mail host generating a random message for signing. Here
is the quote from the description:
"When a sender wishes to send a piece of E-Mail to a mail host, it
begins communications to the mail host by sending its public key and a
signed piece of information to the mail host for verification. The
mail host first checks to see if the public key is on the authorized
white list of the end users mailbox. If the public key is authorized
by the mailbox, the mail host then verifies the signed piece of
information with the public key to see if the information was
correctly signed. If it had been correctly signed, the mail host
allows the sender to send its message to the mailbox."
Note that there are some omissions from this text. Primarily, what is
the piece of info that the sender is signing, and how does the sender
get it? I was assuming that because this is an unsoliciting email by
the sender, it has not (necessarily) had any prior communication with
the mail host, and therefore invents its own info to sign. It would
also need to send this info along with the signed version of it.
Are you saying that when the sender uses a public key/signed info
method of connecting that there is some prior communication with the
[recipient] mail host whereby the mail host provides the data to sign
? If that's the case, then that is a vital piece of information
missing from the protocol description.
Second of all, the signed piece of information isn't distributed. Once
identity is verified the information is discarded so the only way for a
spammer to gather signed information would be to be a mail host. Aside
from the fact that gathering signed information has no benefit.
I think this goes back to the method for choosing data to sign. If it
is on a per-recipient basis, and decided by the recipient mail host,
then yes I agree with you. However, that is not stated in the protocol.
David Finnie
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg