William F. Friedman - Solution of Cipher Machines

Solution of Cipher Machines

During the 1920s, several new cipher machines were developed generally based on using typewriter mechanics and basic electrical circuitry. An early example was the Hebern Rotor Machine, designed in the US in 1915 by Edward Hebern. This system offered such security and simplicity of use that Hebern heavily promoted it to investors.

Friedman realized that the new rotor machines would be important, and devoted some time to analysing Hebern's design. Over a period of years, he developed principles of analysis and discovered several problems common to most rotor-machine designs. Examples of some dangerous features included having rotors step one position with each keypress, and putting the fastest rotor (the one that turns with every keypress) at either end of the rotor series. In this case, by collecting enough ciphertext and applying a standard statistical method known as the kappa test, he showed that he could, albeit with great difficulty, crack any cipher generated by such a machine.

Friedman used his understanding of rotor machines to develop several that were immune to his own attacks. The best of the lot, the SIGABA — which was destined to become the US's highest-security cipher machine in World War II — was co-invented by Frank Rowlett. At least one patent related to it was finally granted after Friedman had died.

In 1939, the Japanese introduced a new cipher machine for their most sensitive diplomatic traffic, replacing an earlier system that SIS referred to as "RED." The new cipher, which SIS called "PURPLE", was different and much more difficult. The Navy's cryptological unit (OP-20-G) and the SIS thought it might be related to earlier Japanese cipher machines, and agreed that SIS would handle the attack on the system. After several months trying to discover underlying patterns in PURPLE ciphertexts, an SIS team led by Friedman and Rowlett, in an extraordinary achievement, figured it out. PURPLE, unlike the German Enigma or the Hebern design, did not use rotors but stepper switches like those in automated telephone exchanges. Leo Rosen of SIS built a machine — as was later discovered, using the identical model of switch that the Japanese designer had chosen.

Thus, by the end of 1940, SIS had constructed an exact analog of the PURPLE machine without ever having seen one. With the duplicate machines and an understanding of PURPLE, SIS could decrypt increasing amounts of Japanese traffic. One such intercept was the message to the Japanese Embassy in Washington, D.C., ordering an end (on December 7, 1941) to negotiations with the US. The message gave a clear indication of impending war, and was to have been delivered to the US State Department only hours prior to the attack on Pearl Harbor. The controversy over whether the US had foreknowledge of the Pearl Harbor attack has roiled well into the 21st century.

In 1941, Friedman was hospitalized with a "nervous breakdown", widely attributed to the mental strain of his work on PURPLE. While he remained in hospital, a four-man team — Abraham Sinkov and Leo Rosen from SIS, and Lt. Prescott Currier and Lt. Robert Weeks from the U.S. Navy's OP-20-G — visited the British establishment at the "Government Code and Cypher School" at Bletchley Park. They gave the British a PURPLE machine, in exchange for details on the design of the Enigma machine and on how the British decrypted the Enigma cipher. However Friedman visited Bletchley Park in April 1943 and played a key role in drawing up the 1943 BRUSA Agreement.