[RC5] Cracking 40 bits in realtime

Dan Ports danz at lanminds.com
Fri Jan 2 05:18:18 EST 1998


On 01/01/1998 2:00 PM, James Mastros (root at jennifer-unix.dyn.ml.org)
wrote:

>I like this idea, oddly enough.  Unfortunately, it hurts civilins along the
>way, so I cannot support it.  Also, it isn't possible without major 
>re-writes:
>1)  We would need to distribute the crypt-text automaticly.
>2)  We would need to have a more complex way of checking if it is a proper
>    decription.  (What way?)

I don't agree at all with the idea of cracking 40-bit encrypted files
on demand, because it could very easily be used for the uses that we
want to avoid by encouraging longer key lengths; however, this brings
up an interesting question. If we don't know what the plaintext is,
how can we determine if we have successfully decrypted the message?
Whereas we know in the DES and RC5 challenges that the plaintext will
begin with "The unknown message is:", if we were cracking a real message,
we probably wouldn't already know the plaintext or a part of it.
Wouldn't this mean that, even if we could test every possible key, we
wouldn't know which one is correct? If this is the case, how could a
brute-force attack like this be applied to an unknown plaintext?


-- 
Dan Ports          http://www.nwlink.com/dalmac/  http://www.netmug.org/
danz at lanminds.com         The main cause of the fall of the Roman Empire
ambclams at netmug.org       was that, lacking zero, they could not express
AmbulatoryClams at IRC       the successful completion of their C programs.

--
To unsubcribe, send 'unsubscribe rc5' to majordomo at llamas.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list