1998-05-13 - [rivest@theory.lcs.mit.edu: Chaffing question]

Header Data

From: Jim Thompson <jim@smallworks.com>
To: cryptography@c2.org
Message Hash: 158e9eb89d35682f4f9deb57cc017325456cc2d285598163abd3ed61a73a1529
Message ID: <199805130747.CAA18984@hosaka.smallworks.com>
Reply To: N/A
UTC Datetime: 1998-05-13 07:48:08 UTC
Raw Date: Wed, 13 May 1998 00:48:08 -0700 (PDT)

Raw message

From: Jim Thompson <jim@smallworks.com>
Date: Wed, 13 May 1998 00:48:08 -0700 (PDT)
To: cryptography@c2.org
Subject: [rivest@theory.lcs.mit.edu: Chaffing question]
Message-ID: <199805130747.CAA18984@hosaka.smallworks.com>
MIME-Version: 1.0
Content-Type: text/plain



With all of the discussion of C&W, and potential expansion, along with
a sideband discussion of same, I decided to ask 'the man' himself.

The '10K email file' used in the example came from the sideband discussion.

Forwarded for your edification.

 ------- Start of forwarded message -------
 Date: Tue, 12 May 1998 18:24:46 -0400
 From: Ron Rivest <rivest@theory.lcs.mit.edu>
 Subject: Chaffing question

 Suppose you can start with a 10KB email file.

 Running a package (all-or-nothing) transform on it adds only a small
 amount to the length---say 128 bits total, or less (basically the length
 of a key in your favorite algorithm).

 You can break the result into 128 one-bit blocks and one 10KB block.  MAC
 each block.

 Add 128 bits of chaff, intermingled with the first 128 blocks from above.

 The final result has 128+128+1 packets, and total length 10KB (the
 original email) + 257*64 bits (for the 257 64-bit MAC values) plus 128
 bits (the chaff bits), and assuming that the sequence numbers are
 implicit rather than explicit, you get a total of around 12.3 KB.  

	 Cheers,
	 Ron

 ------- End of forwarded message -------

-- 
Jim Thompson / Smallworks, Inc. / jim@smallworks.com  
      512 338 0619 phone / 512 338 0625 fax
Real, cheap Hardware RNG:  http://www.fringeware.com/nscd/







Thread