customs searching laptops, demanding passwords

[John Denker]

I quote from
  http://www.washingtonpost.com/wp-dyn/content/article/2008/02/06/AR2008020604763_pf.html
  By Ellen Nakashima
  Washington Post Staff Writer  
  Thursday, February 7, 2008; A01

> The seizure of electronics at U.S. borders has prompted protests from
> travelers who say they now weigh the risk of traveling with sensitive
> or personal information on their laptops, cameras or cellphones. In
> some cases, companies have altered their policies to require
> employees to safeguard corporate secrets by clearing laptop hard
> drives before international travel.
> 
> Today, the Electronic Frontier Foundation and Asian Law Caucus, two
> civil liberties groups in San Francisco, plan to file a lawsuit to
> force the government to disclose its policies on border searches,
> including which rules govern the seizing and copying of the contents
> of electronic devices.

=============

Most of the underlying issue is not new;  a Joe Sharkey article
about customs seizures of laptops appeared in the NY Times back 
on October 24, 2006.  And it has been discussed on this list.
(The news "hook" here is the filing of the lawsuit.)

One wrinkle that was not previously reported is the bit about
customs officers demanding passwords.  That is something I
have thought about, off and on, and the more I think about it 
the more worrisome it seems.

A) Here's one particularly nasty scenario:  Long ago, the traveler
experimented with using an encrypting filesystem, perhaps the 
dm-crypt feature of Linux.  However, he decided it wasn't worth 
the trouble and forgot about it.  This includes forgetting the 
passphrase.  Now he's at the border, and customs is demanding 
the passphrase.   
 -- Just tell us the password.
 -- I forgot.
 -- No you didn't.
 -- Yes I did.
 -- You're lying.
 -- No I'm not.
 -- Yes you are.
 -- No I'm not.
 -- Just tell us the password.
 -- et cetera.

B) Another scenario:  Your employer adopts a policy requiring
you to use a "blank" laptop when traveling, as mentioned in 
the news article.  They also require you to use an encrypting
filesystem, even when not traveling.  They discover that the
easiest way to "blankify" your laptop is to overwrite the IVs
of the encrypting filesystem.  Now any and all passphrases
will fail in the same way:  they all look like "wrong" pass-
phrases.  Now are back to scenario (A), because customs might 
assume you're just lying about the passphrase.

C) Another scenario:  Customs confiscates the laptop.  They 
say that you won't get it back unless/until you give up the 
passphrase.

D) Tangential observation:  If they were being reasonable, they 
would confiscate at most the disk drive, and let you keep the 
rest of the hardware.  But they're under no obligation to be 
reasonable.

E) Remark:  The fundamental problem underlying this whole 
discussion is that the traveler is in a position where he has 
to prove his innocence ... which may not be possible, even if 
he is innocent.

The doctrine of innocent-until-proven-guilty does *not* apply
to customs searches.  Ditto for the doctrine of requiring
probable cause, search warrants, et cetera.

F) A good way (not the easiest way) to "blankify" a laptop
is to remove the hard disk and replace it with a brand-new
obviously-innocuous disk.  (Small, slow disks are very cheap.)
When you get home from your travels, you can undo the switch.

G) It is fun to think about a steganographic filesystem, with
the property that if you mount it with one passphrase you see
one set of files, while if you mount it with another passphrase
you see another set of files.  

The point here is that you give up one passphrase, they never
know if there is a second;  if you give up two passphrases,
they never know if there is a third, et cetera.

Note that we are talking about cryptologically-strong stego
here (as opposed to weak stego which falls into the category
of security-by-obscurity).

>From an information-theory point of view this is perfectly 
straightforward;  solutions have been worked out in connection 
with code division multiplexing.  However, I reckon it would
have serious performance problems when applied to a hard disk.  
If anybody knows how to do this in practice, please speak up!

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