Schneier on Security

Clive Robinson • October 30, 2014 5:35 PM

@ Gerard van Vooren,

Something else that is also not commonly known about the Lorenz machines, is that they were unlike the Enigma designed specifically at Hitler’s request. Some believe that unlike Admiral Karl Donitz Hitler did not trust the Enigma design and suspected it was to simple for high level links. Others believe Hitler had been told about the short message length issue of the Enigma and thus discounted it’s use for high level traffic. What ever the reason, the Germans did know from the Spanish Civil War that the commercial design of the Enigma had been broken by amongst others the French. Which is one of the reasons that the fast rotor on the military Enigma is at the other end when compared to the Enigmas used during the Spanish Civil War. Likewise the British Typex which worked in very much the same way as the Enigma and was used at Bletchly with it’s rotors wired the same way as the Enigma.

What was special about the Lorenz was that it was broken by just two men at Bletchly neither of whom –like the rest of British Intel– had ever seen one, nore pictures or even diagrams or sketches.

It was the Lorenz machine not the Enigma that gave rise to the “Heath Robinson” machine at Bletchly, that Tommy flowers alone and out of his own pocket developed the first thermionic valve (tube) circuits that gave rise to the first working fully electronic computer. The only real recognition Tommy got was a street named after him… the British Establishment did not take kindly to either engineers or scientists for nearly all of the 20th Century, with awards and honours reserved almost exclusivly for “civil service manderins” who rather than get an honest education had elected to do the “Civil Service Exams”…

Clive Robinson • October 31, 2014 1:15 PM

@ Thoth,

There is quite abit more to the Boris Hagelin and Crypto AG story than that article indicates.

The “NSA back door” actually started long prior to the NSA inadvertantly with the C35 “additive coin counting system”.

Back in 1935 William Friedman and colleagues had the responsibility of testing crypto equipment that came in for the attention of the US military etc. They got to see a lot of systems and were very good at analysing them. We now know that the coin counting mechanism had the problem of some strong keys and quite a few weak keys that could be broken fairly quickly, such as around half an hour with pencil and paper.

What Friedman and his colleagues knew was that there was a problem with mechanical cipher systems, which is they are difficult to come up with a secure design, but very easy to copy one that is secure. Thus you don’t want to put your most secure systems into the field as the enemy will quickly get hold of them and if they believe they are more secure than their own will thus copy them. Which gives you the problem that their comms become as secure as yours, which you probably cannot read in a sensible time frame.

The C35 must have seen like a golden oportunity to some one connected to Friedman. Because of the following reasoning,

If you know how to diferentiate the strong and weak keys, you can as you control the KeyMat ensure your troops only use the strong keys. However unless the enemy is your crypto analytic equal or better then they won’t know which are strong and which are weak keys. If however they did know then they would not have a need to copy the design they would have come up with their own. Now having copied the design there was a very high probability that they would encrypt messages under weak keys to the same proportion that they exist in the key map. Thus such messages could be recovered. But… having gained a proportion of plain text messages from the enemy, these could be used to help recover messages under moderate or strong keys.

Thus using the C35/38 only for tactical field/ front line use and issuing only strong key keymat, if your enemy copied the design then you had a way into their tactical traffic, but they did not into yours.

As we know there were quite a few of these Hagelin machines either sold as “war surplus” or new from Hagelin’s company with in effect minor variations. As we know they like Enigma machines ended up all over the Middle East and Africa and the NSA were happily reading them and using techniques developed by the British via the BRUSA –now UKURSA– agrement which gave birth to the Five Eyes.

It’s highly likely that the reason Friedman was used to talk to Hagelin is that he could most believeably tell the story to Hagelin. And Hagelin knowing that his entire business relied on “trust” would have realised that just one or two words from Friedman into the right ears would kill his business deader than last thanksgivings turkey… Thus his compliance with NSA requests was virtually guaranteed. And the more he was complicit the worse it would be, especially when quite a few Middle East nations were known to kill people in other countries for lesser things.

Now you need to ask yourself a question about the speculation on the Israel -Russia, Russia-Iran leaking of the secret, and if there is a more likely explanation.

You need to read Peter Wrights “SpyCatcher” book for some of the background because he indirectly wrote about it.

MI5 for whom Peter and his assistant Tony Sale –of Bletchly Park Rescue fame– had a problem in that they were not getting much needed assistance from either MI6 or GCHQ, even though MI5 relations with the US agencies was better than for the other British IC agencies.

As Peter detailed in the book he and Tony were developing audio eavesdropping equipment for use against embassies, residencies and suspected spys etc. He also mentions that they developed a way to determin the key settings on mechanical cipher equipment simply by sound. If you think about it Enigma machines are not realy susceptible to this but the coin counting mechanism most certainly is.

Peter goes on to mention how thet used this to get way better cooperation from GCHG and gives details of how they deployed the technique against the Egyptian Embassy.

He also details how the Russians knew about the covert pin hole microphones because some one very senior in MI5 leaked the information to them. Thus it’s safe to assume the same person leaked details not just on the audio key establishment but also the information about the Hagelin machine back doors because the leaker knew about them…

Some years ago I got to chatting with Tony Sale about this and after a little while he realised I knew rather more about it than was given in SpyCatcher, and he indirectly acknowledged it. We also chatted a little on how the “NSA Back Door” secret came out when it did, but the MI5/GCHQ audio technique did not. And further how the Iranians had certainty not speculation on the back door. His view is rather more persuasive and some what simpler.

Tony put it down to the Iranians getting their hands on large quantities of CIA and other US IC documentation from raw intercepts and through to diplomatic traffic going back prior to the coup dete by the CIA in 1953 that put the Shar back in control. How did this happen well you need to think back a little to 1979 and the Iranian revolution. It happened so fast that the Iranian Students over ran the American Embassy before they had managed to shred and burn the archives, something that became public knowledge at the time. Over the next couple of decades students slowly rebuilt the documents from the bags of shredded archives. We know this is certainly true because they have in the past not only said they were doing it they also made quite a few of the reconstructed documents available. This is on balance the most likely way the Iranians had the proof to “out” the NSA backdoor, but not the evidence to out the MI5/GCHQ audio method.