Update

Added Stein, Hasenjaeger, Aumann, Weber, Witt, Schultze and Grunsky in German mathematicians in the cryptologic service.

Russian Fish intelligence – A case of quantity over quality

As I’ve mentioned before the internal Soviet radioteletype network was intercepted during the 1930’s and 1940’s by the Germans and postwar by the Americans.

The intercepted plaintext traffic concerned economic and military matters and was of vital importance in finding out what was happening inside the Soviet Union.

However the Russian Fish intelligence was definitely a case of quantity over quality. This is clearly mentioned in several TICOM reports and matches the American assessment during the early cold war period.

Alexis Dettmann, head of cryptanalysis at the German Army’s cryptanalytic centre in the East -Horchleitstelle Ost, says in TICOM DF-112:

‘The monitoring and deciphering of internal radio traffic was not an assignment of army signal intelligence units but necessarily messages of internal networks were solved and worked on. Special offices in the former German army were occupied among other things with the reception of messages of Baudot circuits, the value of the results however belonged in a different sector. Even in the years 1938/39 a relatively simple devise was constructed which made it possible to reproduce directly on typewriters the Baudot messages which in part ware transmitted by high-speed transmitters. The results from the point of view of content in no wise corresponded to the expectations. Of the entire traffic monitored at great expense at best 10% was useful for economic leaders while military-political matters constituted hardly 1%.. The major portion of these messages was like the content of the long distance telephone messages and contained private or business affairs. It was learned that all these circuits were not only monitored and controlled by the NKVD but in many cases were directed by it, and that in all probability the GUP-NKVD was also responsible in large measure for the issue of cryptographic material for internal radio traffic.’

Otto Buggisch, a member of the cipher machine department of the German army’s signal intelligence agency, gives the same percentage in TICOM I-58:

‘Further on Russian Baudot – B. says that one Dipl. Ing. Gramberg came to group IV with him from In 7/VI (Army Signal Intelligence) and was used to translate the intercepted clear text in Russian Baudot. ‘’ 90% of it was unimportant’’.

The relative lack of importance of each individual message was also recognized by the Americans. According to NSA history ‘The Invisible Cryptologists: African-Americans, WWII to 1956’:

‘The ASA. effort to exploit Russian plaintext traffic began in 1946 with the part-time assignment of several linguists to the target. At that time, however, the Agency's emphasis was on the translation of encrypted messages, and the employment of scarce Russian linguists on plain text was judged to be unwarranted. Later, in May 1947, the effort was revised at the Pentagon. Individuals without security clearances or with partial clearances would sift through volumes of messages and translate all or parts of those determined to have intelligence value. Placed in charge of this group was Jacob Gurin, an ASA Russian linguist who had immigrated to the U.S. with his parents at the age of three.

……………………………………………………………………………………………………………………………………………

From the Agency's inception under William Friedman, its business was the breaking of codes and ciphers. Once the underlying text was revealed, individual messages were translated, and, after a reporting mission was established, selected ones were published on 3" x 5" cards. While individual decrypted messages could be extremely valuable, plaintext messages were most often preformatted status reports that were insignificant when considered singly. Jack Gurin was convinced that if these messages were assembled and analyzed in the aggregate, they could yield valuable information on Soviet defense capabilities.’

 

For both the Germans and Americans the limited value of single messages was leveraged by the huge intercept volumes.

FMS P-038 ‘German Radio intelligence’ says: ‘At the experimental station the volume of recordings, which were made available to the cryptanalysis and evaluation sections of the Armed Forces Cryptographic Branch and the Evaluation Control Center of OKH, averaged ten million transmissions a day.’

Information on the  Anglo-American interception is available in NSA history ‘On Watch: Profiles from the National Security Agency’s past 40 years’:

‘In addition to manual Morse, the Soviets were using a good deal of [redacted] among others. The Soviet plaintext problem was a SIGINT success story from the beginning, from the design of electro-mechanical processing equipment that could handle each new Soviet development to the painstaking analysis of the intercepted communications. A joint American-British effort against these communications in the nineteen-forties led to high intercept volume and new engineering challenges in the face of proliferating Soviet [redacted] techniques.

At one time the United States and Britain together were processing as many as two million plaintext messages a month, messages containing everything from money orders to birthday greetings. The production task was awesome, with analysts manually leafing through mountains of page copy, meticulously screening millions of messages. [redacted] The investment paid off, leading, to an encyclopedic knowledge of what was going on in the Soviet Union. Over 95 percent of what the United States knew about Soviet weaponry in the nineteen-forties came from analysis of plaintext radioprinter traffic. Almost everything American policy makers learned about the Soviet nuclear energy and nuclear weapons programs came from [redacted] radioprinter traffic, the result of fitting together thousands of tiny, selected pieces of the jig saw puzzle.’

 

Update

Added information and made some changes in A who’s who of German signals intelligence.

Compromise of OSS codes in WWII – Part 2

In my piece on the compromise of OSS codes during WWII it was stated that Allen Dulles occasionally used diplomatic ciphers when his own systems where overloaded.

In 1943 the Germans were apparently able to read his messages enciphered on the M-138-A strip cipher. The question is whether this was an OSS strip set or the special set used by the embassy in Berne for diplomatic traffic.

Report SRH-366 ‘History of Army Strip Cipher devices’ says that the Army Signal Intelligence agency provided M-138 strips for OSS use in 1944.

 

This would mean that the system exploited by the Germans in 1943 was probably the diplomatic strip.

Update

Added information on the Christie suspension and made some corrections in WWII Myths - T-34 Best Tank of the war.

Insecure traffic of Soviet GHQ units

During WWII the radio traffic of Soviet units was one of the most reliable sources of information for the German Command. Through traffic analysis and D/F the numbers and location of units could be identified. In cases where the messages themselves could be decoded the Germans could anticipate enemy attacks.

In the first years of the war in the East the Germans could read practically all the Soviet codes. In the period 1943-45 however the SU upgraded its cryptologic security. The top level 5-figure code was enciphered almost exclusively with one time pad and the insecure 4-figure codes of the OKK type were replaced with SUV tables.

This meant that the work of the Germans codebreakers became much harder. However they were helped in their work by a serious error in the Soviet Union’s radio security. Special units controlled by the Soviet High Command (assault, engineers, artillery, supply) did not follow the strict protocols of the standard military formations nor did they use secure codes. These errors allowed the Germans to circumvent the new Soviet procedures.

By monitoring the traffic of the GHQ units assigned to large Soviet formations their concentrations and movements could be followed.

 

 

Source: FMS P-038 ‘German radio Intelligence’

Swedish Army codes and Aussenstelle Halden

During WWII Sweden was neutral but maintained close economic relations with Germany. The German signal intelligence agencies were interested in Swedish communications and they tried to solve their diplomatic and military systems.

Diplomatic systems

The Swedish diplomatic traffic was mainly enciphered with Hagelin cipher machines. The Germans analyzed the traffic but according to postwar reports could not solve it (although one message of 5.000 words may have been solved).

The Allies also targeted Swedish Hagelin traffic and had some success, mainly through physical compromise, but according to a report dated August 1944 (Fish notes report 102) ‘the keys have not been broken since January 1942 and none of this traffic has been read since June of that year’.

Military systems

The military traffic was intercepted and decoded successfully by a unit in Halden, Norway. This was outstation Halden (Aussenstelle Halden). This unit belonged to Feste 9 (Feste Nachrichten Aufklärungsstelle -Stationary Intercept Company) but was attached to the Halden Police battalion for administrative purposes. It was commanded by Lieutenant Thielcke.

The systems solved by the Germans were:

1). SC2 - Slidex type system, read in May ’43.

2). SC3 - 3-letter field code without reciphering, read in April ’43.

3). SC4 - 3-letter alphabetical code without reciphering, read in June ’43.

4). SRA1 and SRA5 - Grille/Stencil systems. First broken in the spring or summer of ’43.

5). SM-1 (Schwedische Maschine 1) - version of the Hagelin C-38. This was solved on operator mistakes and ‘depths’. Some details are given by Luzius, an expert on Hagelin cipher machines at the German army’s signal intelligence agency:

‘7. He was then asked whether they had achieved any other successes with this type of machine. He recalled that the Hagelin had been used by the Swedes, in a form known as BC-38. This was similar to the M-209, but with the additional security feature that, whereas with the American machine in the zero position A = Z, B = Y, etc., In the Swedish machine the relationship between these alphabets could be changed. He could not remember whether it had changed daily or for each message. He himself had worked on this machine and had solved a few messages. It had been an unimportant sideline, and he could not remember details; he thought that it had been done by the same method, when two messages occurred with the same indicators. This had only happened very rarely.

The report E-Bericht 7/44 of Feste 9 has some information on Swedish systems:

The people of Aussenstelle Halden were not successful with all the Swedish codes. According to ‘European Axis signals intelligence’ vol4 the high level grille HCA and the ‘large’ Hagelin (probably a version of the Hagelin B-211) were not solved.

The solution of the tactical codes and the C-38 allowed the Germans to build up the Swedish army’s OOB. Why were the Germans so interested in the army’s dispositions? It seems that in 1943 they contemplated an attack on Sweden.

Sources: European Axis signals intelligence’ vol4, CSDIC/CMF/Y 40 - 'First Detailed Interrogation Report on Barthel Thomas’, TICOM reports I-55, I-64, I-211, ‘Hitler’s war’, E-Bericht Feste 9 - 7/44

Update

I added information on Emile Bollaert, Pierre Brossolette and Forest Yeo-Thomas in German counterintelligence operations in occupied France.