Re: FV Demonstrates Fatal Flaw in Software Encryption of Credit Cards

Nathaniel Borenstein <nsb@nsb.fv.com> wrote:
>> Programs needing secure entry create a "secure entry field" which is
>> really just an imagemap with the digits (and alphas if required) placed
>> randomly about.  The user then uses the mouse to click on these numerals.
>> Ideally the graphics that represent the numerals would be drawn from a
>> random pool and are misformed to thwart any OCR attempts. The graphics 
could
>> be made even more difficult to OCR by mixing in words and pictures to
>> represent the numbers. 
>If any particular program for doing this came into widespread use, we
>could engineer an attack, similar to our keystroke attack, based on the
> specific properties of the approach used.

You could try but I don't think you would succeed.  I have problems doing 
OCR on faxes with a top of the line OCR program.  Don't tell me your 
trojan horse is going to be able to OCR images that are designed to be 
hard to OCR.

Here is an example of an imagemap for secure number entry.

http://www.l0pht.com/~weld/numbers.html

Since this is inherently a visual thing, I thought I would cook up a 
graphic on the web siince you cannot do this via email easily.

      Weld Pond   -  weld@l0pht.com      -     http://www.l0pht.com/
      L  0  p  h  t    H  e  a  v  y    I  n  d  u  s  t  r  i  e  s         
      Technical archives for the people  -  Bio/Electro/Crypto/Radio

      L0pht Open House 2/3/96 at 8:00pm - Live on irc #l0pht - write
      root@l0pht.com for details.

Excerpts from mail.cypherpunks: 30-Jan-96 Re: FV Demonstrates Fatal F..
Weld Pond@l0pht.com (1503*)

> Here is an example of an imagemap for secure number entry.

> http://www.l0pht.com/~weld/numbers.html

I *really* like this example.  That's because it demonstrates so clearly
the security/usability tradeoff that I keep trying to hammer home to
people.

Yes, with something like this -- and a LOT of variation, so it wasn't
the same every time -- you could avoid an attack like ours.  But you'd
also have a user interface that was virtually unusable.  The focus of
the attack we outlined was one particular, naive approach to Internet
commerce that sacrificed a lot of security for usability.  If the net
result of what we've done is to force them to find a better balance, it
was well worth the effort.

Or, to put it another way, I'm not too worried about competing with
software-encrypted credit card numbers if they use an imagemap technique
like the one you've outlined.
--------
Nathaniel Borenstein <nsb@fv.com>
Chief Scientist, First Virtual Holdings
FAQ & PGP key: nsb+faq@nsb.fv.com

The "keyboard sniffer" of FV is really troublesome, and the
extension of this threat will hamper the Internet Commerce
tremendously, I believe. The thing that might have made it
hard to accept the threat for cypherpunkers is that it was 
presented together with a plug for the FV scheme, (which may 
or may not be valid btw.)

But more generally, I see the following happening.

The factors that now are "harmonizing" are;
* the tremendous growth of Inet commerce; Digicash, encrypted
  CardNo's etc. Many of the now proposed schemes have no
  independant "evidence" mechanism, whereby you can settle
  a disputed transaction fairly. You will have to choose
  to believe one of the parts, and that is very often the
  service provider/bank/card company.

* The decline of the "ordinary" card fraud market,
  VISA/Europay/Mastercard is rapidly finishing their
  forthcoming smart card systems. I'd guess this "market"
  is gone within 2-3 years. Some "big organisations" might
  start to move into the new "fraud markets" soon.

* The fact that the PC are such an extremely used platform,
  and that the need for back compatibility will make it
  almost impossibe to add substantial security to it now.

* The fact that anti-virus tools haven't been able to
  eradicate the virii problem even before the "forthcoming
  surge" in virus writing that I believe will come. According
  to a survey by Information Week (Nov 27 -95) 67% of the 
  companies had been hit by a virus the last year, and 12% 
  of the companies had suffered financial loss caused of it. 
  (1293 companies surveyed). 
  Admittedly there are social problems behind the continued spread
  of virii too, but that alone doesn't make them go away. Take
  a look at the article "Virus Authors strike Back" by Alan
  Solomon in "Computers and Security" 11 (1992) 602-606. The
  state of anti-virus tools seemed to be in a rather sad state
  back then, and I really wonder whether they are any better
  now.

* The knowledge about how to write virii has been spread
  rather far - a college kid can get his hands on one of
  the polymorphic virus generators, and start to output
  new self-encrypting virii with the same action routine
  regularly. Also, note that this new kind of virii ("virii
  with a mission") would start to cost immediately, in 
  contrast with the "old kind" that only cost when you 
  have to clean them out, or if they wipe un-backuped data.
  (your fault - core dumped)

* All PC's will be net-connected... Embed a public key in the
  virus, let it encrypt the loot and post it to Usenet
  in the group junk.erotica. You can then harvest the group
  with the secret key anywhere in the world.
  (Be generous, let the virus go away automatically if it
  has "contributed" enough money.)

The pay-off of continously updating your virus to cope with
new protection mechanisms would be enormous. Lets assume that I
employ 10 programmers 2 years from now, that writes new action 
routines and develop new virus types... I bet I could get 
a decent living quite soon. Also assume I settle down in a 
suitable country with lax enough laws, do you believe that I
would be a criminal then? What is the legal status of virii,
and what is this concept of "electronic money" anyway? :-)

I promise, I wont do that. It's not a bet.

I am rather shocked that after wading through hundreds of msgs of abuse
of Nathaniel and FV I haven't seen one message of support; but perhaps I
missed it.

I agree that the original post seemed very self-serving and was poorly
worded for this audience.  However, that doesn't excuse people from
reading it carefully and thinking about the implications.  FV has argued
time and again that their basic strength is that CC number aren't
available for systematic secretive purloinage.  The concern about
collecting CC# on a large scale is one of the argument given for the
importance of using encryption throughout the internet.  Otherwise
tapping the internet backbone has much the same properties.  I thought
we had already agreed that dumpster diving is a fundamentally smaller
threat to the CC system than backbone tapping.  FV is just pointing out
that another systematic weakness exists in the CC/internet scheme.

Whether this is a new, serious concern for internet commerce seems to be
a useful and important topic for discussion.  Fortunately, the
discussion, acrimonious as it has been has produced fruit.  It looks
like Weld Pond's suggestion of using a random imagemap is an effective
antidote.  Making the attack harder by an order of magnitude at least.

Ted Anderson