2013/02/23

Military cryptology; one-time pads and my new HDD

.
Cryptology is a field of great promise, great embarrassment and at times great effort. Some of the first computers were developed to break codes and their value was extreme. Yet, generations later we learnt that the U.S. military doesn't even encrypt video feeds of many of its drones, and this wasn't about the tiny ones. The budgets spent on encryption and deciphering as well as analysing messages can be quite outrageous; the NSA's budget approaches the entire German defence budget.

There are no doubt great expectations for the value of code cracking in the future, but I suppose this value depends on the others' failure.


I bought a 3 Terabyte hard disk drive recently. 3 TB - that's 30,000 chunks of 100 megabytes. 100 megabytes is easily enough for two hours of voice comm, and for much more with higher compression, reduced quality, reduced frequency range and lots of breaks between talking.
This single, portable hard disk drive can hold enough data to immunise me perfectly against code cracking by all computing power of the world - for hours of voice comm with 30,000 different possible voice comm network participants each.
Immunise perfectly? Yes, with one-time pads.
Such one-time pads don't use an algorithm that could be cracked, nor any pattern that could be cracked; they're random. The main problems: 

(1) Both participants of the communication need to have the identical one-time code table (the 100 MB chunk, for example). They need to have met previously or some courier needs to have distributed the one-time pads.
 
(2) Furthermore, you cannot send the very same information to multiple recipients using the respective pads, for otherwise this might enable code cracking.
 
(3) Third problem; no multiple participants (~phone conference), unless they possess the very same one-time pad.
 
(4) Fourth problem; random means random. No exploitable pattern is tolerable in the randomised creation of the pads (this is feasible today with dedicated hardware).
 
(5) A man-in-the-middle attack is possible if the clear message can be guessed or is known to the attacker (usable for disinformation).
 
(6) The one-time pad must be used only once, and you need at least one bit of one-time pad to encrypt one bit of message.

So with a properly created and distributed one-time pad I could maintain voice comm to any of ten thousands of network participants for hours. I could ask a third network member to generate a new one-time pad for me and the participant I'm communicating with if our shared one-time pad nears its end. He could tell us the new one-time pad in encrypted comm, so we can keep chatting securely. 
The NSA would have no chance to decipher this communication unless we made some mistake. They could get double their budget, another huge supercomputer, yet another, a couple even bigger ones - and would still fail with brute computing force.


The data storage required to enable thousands of possible participants to talk to each other was the prohibitive problem of one-time pads for generations; that's why the extreme ease of storing data is so important.
It has been calculated that all printing presses of Germany would have been required to print the one-time pads required for the encrypted communication of World War Two (or so I've read somewhere). This was utterly impractical, and thus they used mostly a machine (algorithm) and a common code of the day for many participants at the same time. They only needed to distribute the super-short info about the day's code, and could even update this through the encrypted messaging.
It was their bad luck that the algorithm wasn't complicated enough to prevent a computerised and timely decryption.


We don't need to use algorithms for the truly secret messages any more. We can use one-time code tables nowadays. It's still OK to use the simpler algorithms in many applications and also during peacetime in general, but I wouldn't expect a competent (= important) opponent to make use of encryption algorithms during wartime or for the most important communications.
Then again I tend to overestimate the opposing team's prowess, so maybe huge budgets for deciphering and analysing can still be justified.

S Ortmann

P.S.: This post existed months ago already, but became a victim of the new blogger software / interface and disappeared during the writing process.
.

No comments:

Post a Comment