[cryptography] Detecting Crypto Compromises

ianG iang at iang.org
Fri Mar 30 22:59:44 EDT 2012

On 31/03/12 13:23 PM, Landon Hurley wrote:
> Hash: SHA512
> Does anyone have any knowledge of academic papers focused on methods
> of detecting whether a crypto scheme has been compromised in situ or
> on how to utilize intelligence gleaned from compromised cipher texts
> without giving away that compromise to the enemy?
> I'm thinking in terms of scenarios like how could the Nazis have
> methodologically shown Enigma's compromise in a systematic manner; the
> converse as well though: has there been research into scenarios
> similar to the Allies and Enigma (i.e. how to not give the game away),
> or has it all just been highly intuitive guesswork? It doesn't have to
> be period sensitive, anything from Caesar to the recent would be helpful.

This is all heavily studied inside the intelligence agencies.  But I 
never heard of it being published in an academic sense, because any 
academic writings would immediately be classified.  It was in a sense 
the biggest meta-secret of the war(s).

There are lots and lots of spy/war novels about this sort of deception 
planning, and plenty of WWII documentaries that reveal the deception 
planning that went on.  An awful lot of it was to hide the use of Enigma 
decrypts.  Some also for the location & dates of D-Day.  Huge resources 
were spent on these exercises, like Patton's mythical 3rd Army and the 
bombers used to invade Pas de Calais.

("Deception Plan" is a formal term of art in military planning, might 
make a good search term.)

(Probably the place to look is declassified documents that are after 
their 50 year timespan.)

Oh, one historical reference (might appeal to Americans): the reason the 
Battle of the Bulge was a surprise attack was that Hitler was pissed off 
at his prior failures, and personally suspected the communications 
channels were leaking his secrets, so all the orders were sent by 
motor-cycle couriers.  E.g., Hitler was right.  His generals were wrong. 
  (This seemed to happen often enough to keep Hitler in power...)


More information about the cryptography mailing list