once more, with feeling.

[Peter Gutmann]

Dirk-Willem van Gulik <dirkx at webweaving.org> writes:

>As to technical options to accomplish this

The mechanisms for this actually already exist, they're just not used.  First
of all, you need to admit that you have a problem: SSL certs by themselves are
more or less useless in providing assurance, the phishers are buying genuine
CA-issued certs for soundalike domains using stolen credit cards just as
easily as legit sites are buying certs for genuine domains, and no amount of
signature checking and CRLs and other PKI paraphernalia will fix that.  As
Matt Blaze once said, "A commercial CA will protect you from fraud by anyone
whose money it refuses to accept", which was meant tongue-in-cheek at the time
but given current practice by phishers that's exactly what a commercial CA
will do, and no more.

It's only once developers have accepted this that you can start looking for a
real solution:

- Use TLS-PSK, which performs mutual auth of client and server without ever
  communicating the password.  This vastly complicated phishing since the
  phisher has to prove advance knowledge of your credentials in order to
  obtain your credentials (there are a pile of nitpicks that people will come
  up with for this, I can send you a link to a longer writeup that addresses
  them if you insist, I just don't want to type in pages of stuff here).

- Implement key-continuity at the client, i.e. warn if submitting a password
  or credit card number to a site you've never visited before; display the
  password in cleartext if submitting over a non-SSL connection (this is
  better than any explicit warning); (insert standard list, I have a whole
  pile of these, including responses to objections).

- Use a service like Perspectives,
  http://www.cs.cmu.edu/~perspectives/index.html, to catch here-today-gone-
  tomorrow phishing sites (this is actually a service that someone like Google
  could provide, they crawl the web anyway so keeping track of how long a
  server key's been in use should be a relatively minor change to the existing
  process.  Anyone from Google security reading this list?).

- ... and many more, I won't list them all here.

So it's a two-step process, the first of which is to admit that you have a
problem.  As EV certs have shown, we haven't really even reached that first
step yet.

Peter.

---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com