Other Directory Sites: SeekWonder | Directory Owners Forum
The following archive was created by hippie-mail 7.98617-22 on Fri Aug 21 1998 - 17:16:20 ADT
Marshall Clow (mclow@owl.csusm.edu)
Fri, 27 Mar 1998 09:14:09 -0800
cmcmanis@freegate.com wrote:
>Wouldn't winnowing and chaffing be worthless if your adversary had
>access to the entire message stream.
>
>Postulate: Given that Charles has access to all message traffic
>between Bob and Alice, by recombination of packets with unique
>serial numbers, and the ability to recognize the full message (which
>would be easier with a packaged message) the problem reduces to finding
>the combinatorial set of packets and testing each resulting 'message.'
>
>Given the abilities of an even moderately powerful machine, it would
>seem that unless the total chaff and wheat exceeded something on the
>order of 100mbytes the message could be recovered in a relatively short
>time.
>
You are correct, except for your estimates of the amount of work required.
Suppose that the message was 1024 bytes long (a short message!), sent
one bit at a time, and there were two streams, which were the complement of
each other. That means that you would have 2^^8192, or about 10^^2400
possible messages to test and decrypt. Checking a trillion messages a second,
it would only take you several lifetimes of the universe.
Interestingly enough, this imples that you would generate every single
possible 1024-byte message as part of your testing. How could you
determine from examining the message which is the correct one?
"Bring the drugs to 1234 Main Street at 12:30 AM 03/31/98"
will be one possible decryption, as will
"We will meet in the rectory for choir practice at 7:30. Everyone is welcome"
Anyone see any holes in this analysis?
-- Marshall
Marshall Clow Adobe Systems <mailto:mclow@mailhost2.csusm.edu>
Warning: Objects in calendar are closer than they appear.
The following archive was created by hippie-mail 7.98617-22 on Fri Aug 21 1998 - 17:16:20 ADT