Re: Other SMTP verification methods

Yakov Shafranovich <[email protected]> wrote:
> All of these callback schemes have possible DDOS and privacy problems, 
> not to say they should not be done but we must tread carefully.

Callbacks and message signatures do the same thing, except that
callbacks do it much, much, worse.  In both cases the recipient asks
the putative sender whether they actually sent the message.

Signatures are forward security, the sender puts the signature on the
message, and it's there for the recipient to check.  Callbacks are
backward security, the recipient only retrieves the verification data
after it's received the messages.

Even if we posit a callback scheme using some channel other than SMTP
to the sender, so it doesn't annoy and mailbomb senders, it still
requires substantial extra traffic for each message.  To avoid
spoofing attacks (bad guy replaces contents of legit message with spam
and forwards it along) the senders have to remember something about
all of the messages they sent for as long as verification requests are
likely to come in.

This is the real reason I have no interest in looking at callbacks,
and I'd much rather put the effort into signatures.

Regards,
John Levine, [email protected], Taughannock Networks, Trumansburg NY
http://www.taugh.com