On Wed, Feb 09, 2011 at 05:18:35PM -0800, Derek Simkowiak wrote:
>  TrueCrypt has a plausible deniability feature for such countries:

The protection of encrypted data in opposition to legal obligation or
duress has long been a problem. Although I know of no implementation
for encrypted disk systems, the (m,n)-threshold scheme could be used
to provide a solution.[1]  In this system, m people hold partial keys
and if n of them provide a partial key, the encryption/decryption can
be performed.  Here's the interesting part: If m-n+1 people decline or
PROVIDE AN INVALID PARTIAL KEY, the decryption will fail and IT'S NOT
POSSIBLE TO DETERMINE WHO PROVIDED INVALID PARTIAL KEYS. A practical
implementation would use a symmetric key for the data encryption and
decryption, but the key is available only after it's been decrypted
by n keyholders.

Sigh, I know trojan code could reveal the symmetric key. You gotta trust
the system unless this was implemented as part of the storage system's
firmware. Furthermore, it would be a great pain to have gather n people
to boot the system. I just couldn't pass up the opportunity to share
this unwieldy, but perhaps sufficient solution. Oh, and you tell the
judge "I didn't implement this to deny you access. I'm fearful some
bad guys may threaten us and now only a small number of keyholders
are sufficient to deny them access in secret."

[1] Bruce Schneier's Applied Cryptography describes in the discussion
of "Secret Sharing".

-- 
Randolph Bentson
[email protected]

Reply via email to