Organisation of cryptanalytic effort against Soviet Union Review of Central Office organisation Prior to 1939, the
Intercept Control Station [Ref 5.1] had a section for handling Russian traffic, but little was known of its operation or achievements. Early successes against Russian cryptography were evidenced by
cryptanalysis of 2,3, and 4-figure codes with 5-figure codes broken at the start of the war using a Russian 5-figure code book obtained from Finland, that was used in the first year of the war. When HLS was replaced in 1941 with the new agencies:
OKH/Chi and HLS [Ref 5.2], the Soviet evaluation section was moved to LNA and the cryptanalysis unit for the Soviet sector remained with
Inspectorate 7/VI. In late 1941, on a recommendation by
Kunibert Randewig, both cryptanalysis and evaluation sections were moved to
Loetzen,
East Prussia. This section formed the nucleus for the third central agency, the HLS Ost [Ref 5.5]. From that time, until 1944, the German Army Signals Intelligence activities were sharply divided into non-Russian, which were undertaken at
Inspectorate 7/VI and into Russian activities, performed by HLS Ost and the LNA. In October 1944, the three agencies were amalgamated together into the GdNA, into a central agency. For a breakdown of Russian signals intelligence activities, Group II under Captain () Gorzolla and Group IV under Major Rudolf Hentze were the principal units with the GdNA which worked on Russian systems [Ref 5.6.1].
Review of Field Office organisation Cryptanalytic effort against Soviet Union Introduction The Soviet Union state used various military cryptographic systems in order to secure its communication. Two organisations prepared and evaluated cipher procedures. This was the
NKVD 5th Department and the
Red Army's 8th Department of the main intelligence directorate
GRU. The soviet military used 2,3,4, and 5-digit codes enciphered with substitution methods or with additive sequences used with the most important 4 and 5 figure codes. The NKVD relied on figure codes enciphered both with substitution and addition methods. Partisan groups also used figure codes enciphered with additive sequences or transposed based on a key word. The
diplomatic service used 4-digit codebooks enciphered with
One-time pads tables.
2-Digit codes The
Soviet Union 2-digit codes were used by the
Russian Army,
Air Force and
NKVD. In the Army, they were used by Army Groups, Armies, Corps, Divisions, and Regiments; and by small independent special units such as Combat Engineer Brigades, Motor Regiments, and artillery. In the NKVD, they were used on regimental communication close to the front, and from divisional level downwards. Cryptanalysis of 2-digit systems was done mostly in
FAK units at the company level, but was also handled by
NAA, the
NAAS and at HQ level of the GDNA itself. The recipher was carried out using the 10 x 10 Latin Square, i.e. no figure repeated in any one column or line. A line or column of this square was allotted to each day of the month. Ex: 7th, 12th, 19, 18 May Line A 1st, 9th, 13, 29 May Line G 3rd, 6th, 10, 19 May Column E The recipher of the bigrams 17, 86, 00, 80 on 7 May would be according to Figure I be 10, 54, 99, 59. It was clear from interrogations that 2-digit codes were not always in use, but were being constantly read. POWS of
NAA 11 stated that the last known table was PT-43, and it could not be constructed. For Lieutenant Alexis Dettman, who stated that solving PT codes was: :
merely a form of crossword puzzle Dr. Wilhelm Gerlich of NAAS 1, stated that one or two men at
NAAS 1 were able to cope with decipherment of 2-digit messages, especially as the greater part of them were being solved at company level.
KONA 1 prisoners stated categorically that the Soviet Union has ceased using 2-digit codes after 1943, however other evidence would seem to have indicated that they were used up until the end of the war by the Red Army and NKVD, particularly by units in active combat, and were still being actively used, even if not appearing.
3-Digit codes 3-digit codes were used by the Red Army, the Air Force and the NKVD. They were first used in 1941–1942, initially mainly by the Air Force and later, more widely, by the Army. Traffic from Army Group down to Battalion had its own 3-digit code. The 3-digit codes were replaced by 3-digit Signal Codes in 1943, which were used by all units. For the NKVD, they were used by the
Black Sea Fleet and from division downwards. The solution of these codes was done mostly in
FAK companies, but also handled by the NAA, NAAS and the HLS Ost. They were first noticed in February, 1941, and used increasingly from May 1941 at the beginning of the
Russian campaign (
Eastern Front (World War II)). From then until late summer 1942, the Air Force made the most use of the code and each Air Division had its own cipher. In 1942, the
48th Army, the first Army to use it, started by using a 3-digit syballic code. Interrogations of personnel from Feste 10 and KONA 1:
By the time of Stalingrad practically every Army engaged in the battle had its own 3-digit cipher. Dettman and Samsonov do not discuss any 3-digit codes as such, so it must be assumed that they considered these types of codes similar to the 2-digit PT series, and did not warrant discussion as a special sub-type.
Technical description The first 3-digit ciphers were very simple in form, consisting of several pages, ten at most, and contained alphabetically semi-hatted or completely hatted, the numbers 1 to 0 and punctuation marks. See Fig IV. Soon afterwards the letters of the alphabet were put in their alphabetic position in the book. See Fig V. Besides the above types of codes, syllabic codes contained few words but practically all the possible syllables which could be formed out of 2 or 3 letters. All codes of this kind which were broken were alphabetic. See Fig. VII After some months, the strictly alphabetic nature of the books was given up and the alphabeticity only maintained within the area of a letter. The numbers were at the same time distributed at random over the page. See Fig VI. The book could have 1000 groups, but Oberleutnant Schubert stated that the average number of groups was 300–800, stating:
in general small scope, but frequent change. If the codes were smaller than 1000 groups, alternatives were given either to pages or first figures of the lines. The substitutions could be constructed without any recognizable system or they could be made up from a Latin square, with the square usually lasting about one month, although with the Air Force, it sometimes lasted longer. Schubert stated that towards the end of the war, there appeared quite isolated 4 to 7-digit substitution, presumably private systems of the respective cipher departments, but appeared very seldom. From the beginning of 1943, most 3-digit codes no longer contained single letters but used the initial letter of the words for this purpose. In order to indicate which meaning was to be used two groups were introduced: :: 736 1) Read the initial letter :: 737 2) Read the whole word The 3-digit groups were read in the order of page-line. The recipher of 3-digit ciphers was as varied as in the case of 2-figure ciphers. Each figure was re-enciphered separately. This is to say, for the page numbers, a hatted order of the figures 1 to 0 was substituted. Thus Page 1 became 4, 2 became 6 and so on. Fig. VIII. 3-digit code
R3ZC which was mentioned by Corporal A. Faure of
NAA 11 and was discovered as part of the Norway codes. It was a code with 10 pages of 100 positions each, 10 x 10 alphabetically arranged. One column of each page, the 8th was empty. Only the 100's and 10's figures of each group were reciphered. The 3-figure code, it is clear from all interrogations, was mainly used by the Army, but also by the Air Forces. According to a report by Uffz. Karrenberg of Group VI, on regimental networks and for less important messages on the level of assault armies, mainly the 3-digit codes (with a 2-digit Latin square encipherment) was used. KONA 1 personnel stated that every Army Group down to Battalion had its own 3-digit code. A good deal of
plaintext was inserted into the 3-digit enciphered code as it was transmitted. Gerlich pointed out the advantages of plaintext insertion stating:
They often gave words and names not contained in the code. 3-figure traffic was always solved when one encipherment was available 3-figure traffic was only slightly more secure that 2-digit code traffic and was 80% readable. It was known that Air Force codes were often current for much longer periods than those for the Army, and thus be able to find
depth. Army 3-digit codes were changed after a big operation and were in use for one week to one month. However, all 3-digit codes were readable and Lt. Col. Mettig stated that from spring 1943 to 1945, 2-digit and 3-digit traffic was read regularly, providing a large amount of tactical intelligence. The operational 3-digit codes were discontinued in 1943 and replaced by 3-digit Signal Codes, which were different only in that they contained no letters, only words and phrases of importance, and were not alphabetic but had meanings grouped under headings such as
attack,
defend,
enemy movements. Each meaning had two or three, 3-figure groups allotted to it. As in the case of the former code, anything not in the codebook was sent in plaintext. It was assumed that these types of codes were also read as consistently and as fully as the 2-digit and 3-digit codes they superseded.
4-Digit codes 4-digit () codes were used by the Army, Air Force and NKVD. There were used in the Army as General Army Codes, called General Commanders Codes, and on lower operational levels by mobile formations such as Tank and Mechanised Corps, Tank Armies, and Tank Administration and Supply Units. For the NKVD, they were used on Railway and Transport Nets. Solutions to these codes were handled at FAK, NAA and GDNA level. At FAK level, the companies usually had a 4-figure unit, as well as a 2-digit unit and 3-digit units, in a cryptanalysis department. Lt. Schubert of KONA 6 stated under interrogation that:
The Soviet Army keys are 3 or 4-digit systems. The basis is the same. The construction of 4-digit codes was the same as 3-digit codes, except the book had a possible 10,000 groups instead of 1,000. with the book length being described as either six or seven pages each with a block of consecutive numbers, or with KONA 1 personnel stating that book length being of from 5 to 100 pages long. The latter value seemed more likely. The Air Force Codes usually had about 10,000 groups and the Army 5,000 groups, and each page could have a variant page designation.
Technical description The actual construction of pages varied in 4-digit code books as much as it did in 3-digit books, in respect of alphabeticity and sequence of numbers. But the methods of enciphering the last two digits were still more varied than those used in 3-figure systems:substitutions by row, digraph substitution (in comparison to single-letter substitution in the case of 3-digit codes), combinations of these two, abbreviated figures in the substitution, and others. Uffz. Corporal Karrenberg, who was a GDNA cryptographer, and was a specialist on Russian
Baudot stated, while under interrogation, in his discussion of the
Russian Cryptanalysis course given for field training, that 3-digit or 4-digit could also be reciphered on an adder. For this purpose a text chosen at random is enciphered (likewise by the code) and the code text added to or subtracted from (non carrying). There was no other indication in interrogations that this method was met in actual practice. The general method seem to have been variant page designations for the book (2-figures) and encipherment of the last 2-digits by various means, including 2-digit
Latin square. From May 1942, the Air Force began to use 4-digit codes in many different forms, but often of the simplest construction. In the middle of 1943, many mobile formations, Tank and Mechanised Corps, Tank Armies and Administration, and Supply units also started using these 4-digit codes. They were also used by the Railway and Transport Nets. Dettmann and Samsonov described the first
general army and air force code, which was a 4-figure with 4,600 groups, enciphered by digraphic substitution. The successor codes were the names
OKK5 to
OKK8 which were General Commanders Codes that rapidly replaced one another from 1939 to 1941. OKK 5 was captured in the
Winter War, and
OKK 6, 7, 8 later. Dettmann and Samsonov stated: :
All these systems, however, recovered by cryptanalysis before their capture and were made completely and currently readable. Ltn. Harry Loeffler of Feste 10 stated that a: :
General army 4-digit cipher was last observed in use in Tajikistan (North Persia) in the winter of 1941–42. It had 50 pages, each designated by two alternative Bigrams and 100 lines to each page. There was no indication in interrogations that 4-digit codes were either less or more frequently employed. In 1944, 4-digit Signal Codes, of the same form as the 3-digit Signal Codes, appeared, and were used by army groups. and probably superseded in great part, if not wholly, the 4-digit codes discussed above. Four-digit codes were changed less frequently than other codes but Dr Wilhelm Gerlich's of
NAAS 1 of
KONA 1 stated, the change in encipherment was not quick enough to make decipherment impossible [by Germany]. Certainly the 4-digit codes gave Germany cryptanalysts a certain amount of trouble, as a large amount of material was
absolutely necessary and the majority of unsolved 4-digit codes and ciphers were abandoned because of an insufficient number of messages. Captain Holetzlo, a member of
LN Regiment 353, the Luftwaffe signals regiment intercepting Russian Air Force traffic, speaking mainly about 4/F air force codes (ground/ground) stated that they were
only 60% readable. Corporal Heudorf of
NAA 8 of
KONA 2 stated that later 4/F traffic provided some difficulty in cryptanalysis, but an Engineer unit's in March and April 1945 was being read concurrently. In the interrogations of KONA 1 members, the following 4/F codes were listed as solved: • 4-digit code of VI Guards Mech. Corps of
1st Ukrainian Front from January 1945 to end of war. • 4-digit code of 152 Tank Brigade of
60th Army of the 1st Ukrainian Front. • 4-digit code of Tank Supply and Administration Authorities of the 1st Ukrainian Front. • 4-digit code of the 76th Regional Air Base (Russian 76 RAB). • 4-digit code of Supply Units of
13th Army of 1st Ukrainian Front. • 4-digit code of
3rd Guards Tank Army. Horst Schubert stated: :
We broke Army 3 and 4-digit re-enciphered books. These were enciphered on a conversion table. Early in the war we read most of this traffic, but by the end only 40% to 50%. 5-Digit codes 5-digit codes were used by the Army, Air Force and NKVD. In the Army, they were used by the
People's Commissariat for Defence (), Army Groups, Armies, Corps, Divisions and Brigades. In the Air Force, they were used by Air Armies, Air Corps, Air Divisions, Regional Air Bases, Anti-Aircraft Corps and Divisions and Anti-Aircraft units. They contained strategic, tactical, personnel and supply matters, and political reports and directives. With the exception of a short period in 1943, when KONA 1 did independent 5-digit cryptanalysis, solution of 5-digit codes was handled exclusively by the GDNA. In discussing the achievements of the
OKH/Chi, Lt. Col. Mettig, stated glibly: :
The breaking of the Russian 5-digit reciphered code...was the most outstanding cryptanalytic achievement of In 7/VI. The Soviet 5-digit code was broken chiefly by War Administration Inspector () Lieutenant Alexis Dettman. When rating the relative importance of cryptographic including cryptanalysis achievements contributing to total intelligence, Mettig was most impressed by the continuous breaking of the Russian 5-digit code despite the difficulties that were experienced after spring-summer 1943. However Mettig could have been stressing organisational changes carried out to facilitate the breaking of codes rather than actual cryptographic achievements. Dr
Otto Buggisch, also of
OKH/Chi stated that Mettig had few fundamental ideas of cryptanalysis. Certainly other evidence exists that points conclusively and without doubt to an almost complete failure on the part of the German cryptanalysts to make any progress with the solution of the 5-digit code. Corporal () A. Althans of the KONA 1 stated that cryptanalysis of 5-digit codes was only possible with the following conditions: • There was a number of messages, at least three, which had the same additive applied. • The 5-digit code had been captured. Dettmann and Samsonov stated that codes
011-A,
023-A,
O45-A,
062-A and
091-A used successively from the beginning of World War II to the end, did prove difficult from a pure cryptanalytic solution. They stated: :''It is interesting to point out that during the course of the war all the newly appearing versions of each cipher were captured through fortunate circumstances and always so soon that the originals were almost always in the hands of the cryptanalysts at the instance of their being put into use by the Soviets. In the reports of personnel from
NAA 11, a Captain Schmidt stated that: :
With regard to Russian [Soviet] traffic, the Abteilung [Battalion] did everything up to and including 4-digit. 5-digit they considered insoluble and forward these to the GdNA. One Corporal Karrenberg, of GdNA, while under interrogation discussed 5-letter and 5-digit codes used for operational orders, stated: :
These were so-called Blocknot codes, which were only used once and were therefore unbreakable. Blocknots were random sequences of numbers contained in a book and organised by numbered rows and columns and were used as additives in reciphering and were considered as a type of
One-time pad. The GdNA mathematician
Johannes Marquart conducted research on Blocknots and was unable to determine how they were generated. Corporal Karrenberg stated flatly: 5-figure and 5-digit messages were not touched at all. In general very little work was done on decipherment of 5-group messages, although these contained the most important operation reports. They were only used to identify units and were only read if code books happened to have been captured. In 1943,
KONA 1, for a period, undertook 5-digit
cryptanalysis independently of GdNA, but the general practice was for all units to forward 5-digit traffic directly to Berlin to GdNA for possible cryptanalysis and evaluation. Karrenberg stated: :
even at the GdNA HQ, very little attention was give to 5-digit messages and very little enthusiasm displayed in working on it. Only the preambles were used to identify units, from Blocknots and indicator groups. Finally Lt. Schubert, when questioned on possible success on 5-digit codes replied: :
In the Finnish campaign the book was captured and the Russians used the one-time pads over again. Because of this we had considerable success. Recently the Russians used the [one time] pads correctly and only very few messages were read, these through re-encodements. Technical description This small measure of success was due to the fulfilment of the two conditions set forth by Althans for successful cryptanalysis. Althans stated that successful cryptanalysis is possible only if: • There are a number of messages, at least three, which have had the same additive applied. • The 5-figure code was captured. Cryptanalysis starts from the mathematical rule that the difference between two code groups remains constant if both have the same additive row applied to them. For example: The most important cryptanalytic aid was therefore the catalogue of differences, a numerically arranged table of the differences between the most frequently used code groups. Directly the code, which changed approximately semiannually, was captured, about 1000 frequently used clear groups (positionen) were written out by the Gen. d. N.A., arranged and serially numbered according to frequency, and then each subtracted from the other, non-carrying, by
Hollerith (
Herman Hollerith) machines and entries made on the catalogue as follows: "41909 17-32" which means: In 1943 KONA 1 for a period undertook 5-figure cryptanalysis independently of GdNA. For this purpose a difference catalogue of 200 clear groups (which equals 19900 differences) was compiled by hand. A calculation aid, in the form of a wooden frame with five paper belted wheels I to V, where one wheel represented each digit of the 5-figure group, was used. The apparatus also served as an aid in actual decoding, when a great deal of non-carrying addition or subtraction had to be done. Note: little information is available from TICOM as to how this device functioned. The
Finnish predecessor to the
Finnish Defence Intelligence Agency, during the early period of World War II had captured and turned over to the Germans, a Russian 5-figure book which was used continually until the
Winter War. An additional copy had been captured by the Germans and though the Russians introduced a new 5-figure code on 1 April 1942, the changeover was faulty, and it was possible to establish 2000 groups of the new code within a week. Indeed, it was possible at this time to leave the decoding of this procedure to the forward KONA units. The Russians gradually improved their security, however, and in the spring of 1943, altered the
indicator group system and split up the code into various front sectors. As it result it was necessary to collect all the 5-figure at
Giżycko (Lotzen) intercept station and to use the Hollerith department of the
In 7/VI. Only thus could have the necessary
depth on a day's traffic be achieved. The only description of 3-F address codes was given in the KONA 1 reports.
Technical description 3-Figure (3-Digit, or 3-F) Address codes were used in conjunction with 5-Digit messages to enable wireless stations to ensure that messages arrived at their proper destination. Their construction was similar to that of the PT-39 or PT-42 codes, described above, but they contained only figures, unit designations, authorities, words such as
for or
from. See Fig X. Example of an address: A code of this type was first used at the beginning of 1944 on the 1st and 2nd Ukrainian and the 1st White Russian Fronts for communications between their Army Groups and their respective Armies and independent Corps. In the summer of 1944 armies began to use similar codes with their subordinates units and latterly the use of such codes increased still further. The solution to these codes, given a fair amount of material, was generally easy. They were often simpler than the example given in Fig. X and sometimes appeared in the form of straightforward lists with each meaning numbered consecutively of 100 pages instead of 10. See Fig XI. The same variations in construction occur as with three-figure. • Alphabetic with the letters of the alphabet grouped at the end of the Code. • Each letter of the alphabet placed before the part devoted to words beginning with that letter. • Semi-hatted, i.e. with all the words beginning with one letter grouped together but the groups in random order. • Letters of the alphabet with alternatives. • Syllabic Codes (with words distributed over the code). • Letters of the alphabet with alternatives. :* Read the initial letter. :* Read the word. • With the two groups directly above appearing several times, often on each page. Codes had anything from 5 to 100 pages. The Air Force Codes usually had about 10,000 groups and the Army 5000 or fewer. According to members of KONA 1, the solution of these codes given a fair amount of material, was generally easy. Members of the
NAA 11 stated the following: :
Addresses to personal names rather than titles were common in all Russian traffic and this suggested the use of initials. The vagueness of this answer surprised the interrogator who asked if the addresses, being enciphered on the PT table, were not read currently. The answer was that unless the same address was used frequently and some outside hint was given they were unusually unable to read the address. They supposed it used values which had special local meanings added to the table. Miscellaneous codes Most of the statements given here are taken from the interrogations of members of
KONA 1. It is not the complete story, as the POW's themselves were aware: :
The examples described are only those ciphers whose basic construction was established. There were many types of cipher which were only partially broken (solved) and whose basic form could not be established, these are not mentioned.. The number of Russian ciphers of all types that were broken was about 3000. 4-letter codes Two forms of four-letter codes first appeared in practice traffic between Army Groups and Armies and Independent Corps of the
1st Ukrainian Front in November 1944: • Revolving Stencil : This consisted of a sheet of paper ruled off into 8 x 8 squares. On top was placed another sheet in which 16 holes were cut of corresponding size to the squares on the first sheet. These holes were so cut that if the sheet is turned through 90° round the centre point in the four possible positions and a letter written on the bottom sheet through each hole in each position, all 64 squares on the bottom sheet would be filled. The text of the message to be enciphered was written in the holes horizontally with the stencil in the first then in the second, third and fourth positions The cipher text was then read off horizontally and sent in four letter groups. If a message was longer that 64 letters the process was repeated as many times as necessary. The revolving stencil was changed from time to time. • Transposition :This cipher is a simple
transposition cipher, the key given by a
Keyword and the text being written in vertically according to the key and either upwards or downwards according to accompanying instructions. The cipher text was then read off horizontally and sent in 4 or 5-letter groups. See Fig. XVI The contents of messages sent in both the above types of cipher were usually about technical signal matters though in later stages units and positions were named.
Word-Code From the middle of 1944, a word code was used in the area of the
2nd Ukrainian Front. The book consisted of two halves, each designated by a word such as
SEVER ZAPAD literally meaning North West. Clear groups such as numbers, units, offices, designations such as Tank, Guards, Mechanised, were grouped in two columns. Each half of the code identified was only used by the Army and contained strategic and tactical reports and the names of the units. They were small in size and contained only essential groups. The first half of the code is used with column 1 and the second half with column 2. The order of the halves can be changed and indicated by the indicating word. If the cleargroup consists of more than one word such as
In the former sector = Dolvnost and the wanted word is
former, this can be indicated by saying
the second word of DOLVNOST. See Fig XVII. Example: :: The codes identified were only used by the Army and contained strategic and tactical reports and the names of units. They were small in size and contained only essential groups. Anything that could not be enciphered was sent in
plaintext or in the clear. Coordinates also appeared in messages.
Periodic and Columnar Substitutions Corporal Karrenberg stated in discussing the
German Army Cryptography Course stated: "These rarely appear in Russian cipher systems... Periodic and columnar substitutions concluded the sections on substitutions systems in the course. Not so much time was spent on them as they are rarely encountered in practice."
Coordinate systems These were very varied. Armies made up their own systems and arbitrary reference points and grids were used.
Machine Ciphers The handling of
Teleprinter traffic was processed exclusively at GdNA. Lt. Schubert, whose position at the GdNA chi-stelle was Director of Main unit East, (), stated that teleprinter traffic was worked on in the machine section, specifically Group IV, whose director was Rudolf Hentze. Lt. Schubert thought that messages in
Depth had been read, but was uncertain whether the machine had been recovered. Schubert himself never worked on cipher machines, but knew that the Soviet (Russians) had a machine, already in use at the beginning of the war, but not on military traffic. Corporal Karrenberg spoke of the
Bandwurm, the so-called
Russian FISH, who defined it as Russian
Baudot letter
strip, that should not be confused with Soviet 5-F traffic also carried on
Baudot lines. Karrenberg believed the first traffic was intercepted in 1940 in
Warsaw, and as far as he believed there was no interest taken in it. The first actual knowledge of the traffic, with the same external features, e.g. chat, indicators, was in summer 1943, when the first real interest was taken in it. Dr Pietsch and Döring conducted the initial research. The German intelligence agencies did not capture any of the apparatus used, but felt that it consisted of two parts: • A Baudot
teleprinter. • A cipher attachment consisting of five small wheels driven by one large wheel.
Depths were frequent, but the Germans did not seem to make any attempt to reconstruct the wheel patterns. The system was used by the
Red Army and the
Air Force and to a lesser extent the
NKVD.
Otto Buggisch of the
In 7/VI later GdNA and the
OKW/Chi went into somewhat more historical detail and stated that: • In 1943, Buggisch heard that the
Research Office of the Reich Air Ministry,
Hermann Göring's Research Bureau, had success on a Russian teletype machine and had recreated the action. It was a machine with a very long cycle, not being prime but the product of several small cycles, like the S42. Buggisch did not know the cycles of the individual wheels or any other details. He found this out from Döring, who was undertaking research on the T52. Liaison with the
FA was considered bad anyway. Mettig was particularly opposed to the
SS. Nothing came of the FA work. Buggisch stated in interrogation that the cycle of one wheel was 37, varying from 30 to 80. • Late in 1943, early 1944, OKH began to intercept non-morse 5-impulse traffic, named
Hughes by Buggisch. Group IV worked on it, at the end of 1943, resulting in a
Kompromise and a depth of eight messages with the same settings was created. The section was able to recover 1400 letters of pure key, and to determine the traffic was derived from a 5-figure code, with regular station chat enciphered at the same time on the machine. Corporal Karrenberg stated that part of the depth was created within the same long message, so that the machine had a cycle, at least in this one case of about 1450 letters. The actual number was thought to be very significant by the Germans, as it was prime, so could not be the product of smaller cycles in any way they could image. Germany postulated a tape machine like the T43, or a machine in which the motions of the wheels influenced each other, 1 and 2 affecting 3, 3 affecting 5, etc. as in the T52 (which was known to be insecure). They were not able to prove any theories they had. Germany had chronic personnel problems, and the OKH/GdNA was no exception, being short of mathematicians. Buggisch believed there would be a solution. • After this experience, they devised
Hollerith machinery to locate depths, but in fact only found three or four more cases and none of these gave additional cycle evidence or even furnished another key. • Buggisch thought the traffic slumped, left the unit in June 1944, and the LNA stepped in to improve the solution, to improve reception as they believed the traffic was still present, which it was not. Buggisch stated in passing that their own security idea on the subject of wheel machines of this sort was that the cycle should not be a product of small periods (as in Hagelin cipher machines), even if this was long. Mutual influence of wheels should be used to avoid this, but at the same time care must be taken that too short a period was not created in the process. This in fact had apparently been done by the Soviets, but the fact that it was not repeated suggested to him that they might have seen the weakness and corrected it. Buggisch stressed one fact which had surprised him, that they never had information about either of these machines, from POW's or agent sources. He assumed that the one that the FA broke was not the same because of the difference of cycles. The number of links, according to Corporal Karrenberg, varied according to the number of armies, with the maximum of eight. One end of the link was
Moscow, the other mobile. After 1944, no work was done on the traffic except on the spot. No vital clues to the system were given away by the Soviets, though their security precautions were not considered effective.
NKVD codes NKVD Codes were simple mono-alphabetic
Substitution cipher 2-figure, 3-figure, 4-figure, 5-figure and 5-letter types. They were used without any apparent reason on two large networks: • The networks of the NKVD Central Authorities, that were divided into those of Security Troops, Frontier Troops, and Railway Troops and Convoy Troops. • The network of NKVD Formation, communication between units attached to Army Front Staffs. 4-figure and 5-figure codes were used on the front line: there was a 4-figure code, e.g. used by the military police, and a general 4-figure code used on Staff-Regiment-Battalion links. Solution of NKVD codes was handled by the
NAAS and the GdNA. In interrogation: :
The traffic of the NKVD formed a special group of Russian wireless traffic. The distinction applied equally to the manner of conducting traffic and to the message themselves. The German Army cryptanalysts reflected this Russian
distinction in their own attacks on NKVD systems, allocating the work, again in their own attacks on NKVD systems, to
levels of operation determined by difficulty of solution. Lt Ed. Wöllner of KONA 1, stated that: :
NKVD traffic was always covered, but only by FAK companies. Evaluation and cryptanalysis were done by NAAS. On the same subject Lt. Löffler of
Feste 10 stated that :
all NKVD signals originated in the regiments area were worked on in the NAAS, others were sent to the LNA. But in neither case was there specified what type of NKVD traffic was worked on, and what type was passed to higher echelons. Apparently the distinction corresponded to that observed in the case of actual army traffic. All lower-level operational codes, up to 4-figure codes, could be handled by the NAAS. The 5-figure codes were not dealt with by the
KONA, but handled by 40 to 50 men in the LNA, [Ref 5.4] at
Zeppelin bunker at
Zossen. Löffler thought that a good deal of success was obtained in the case of Far Eastern Traffic. An additive was used for the recipher. Dettman and Samsonov substantiated this point, in their discussion of the German army cryptanalytic effort: :
All the five-place message material from the Army or the Airforce, as well as the NKVD messages, was submitted for the exclusive processing of the GdNA They also gave full descriptions of NKVD systems. There was no discrepancy between the facts in their report and the facts given by Lt. Löffler. The details of description were scattered throughout the report according to types of encipherment. They mentioned the following: • Mono-alphabetic substitution systems common to all NKVD organisations. • Conversion systems: substitution and additive: ::* 1936: 4-figure code: 2500 values, used in district. ::* By 1939: three 4-figure conversion systems with codebooks of up to 5000 values, enciphered by digraphic substitution or conversion table. ::* 1939: first general NKVD 4-figure code, with 10000 book positions enciphered first by single digit substitution, later by additive. At time of capitulation, three 4-figure systems (
ZERNO,
NEVA,
VIZA, see infra) used respectively by: :::* NKVD Troops. :::* NKVD Border defense troops. :::* NKVD Security troops. ::::(200 messages were read daily in all three) :::* 5 figure Railway Troops code, which used 2500 groups with digraphic substitution encipherment. A few 3-figure small codes (read currently) existed, that used letter transposition. These were never found in NKVD traffic number series. ::* 1940: 4-figure single letter conversion encipherment (1941–1942): 4-figure, enciphered by text key, which was letters equalling numbers. ::* 1942: 4-figure enciphered by military technical manual (used by Interior troops) until end of 1944 with encipherment combining single digit conversion and Gama Tables (no description of these is given).
NKVD codes technical description NKVD traffic was always covered, but only by
Long Range Signals companies, with evaluation and cryptanalysis done by the NAAS units. When W/T traffic was restricted during radio silences, NKVD traffic was often the most important source for Radio Intelligence. At such times, it was more completely covered. When the Army W/T was in full swing, coverage of NKVD was correspondingly reduced. On an average day, 6–12 receivers were employed exclusively on NKVD cover. In general the following nets were covered: • Networks of Central Authority. The covered the Chief NKVD authority in
Moscow with the Front HQ's. The NKVD W/T radio intelligence picture, gave indications of groups, or the insertion of new Fronts. E.g. the appearance of a new W/T station communication with the
1st Ukrainian Front and
2nd Ukrainian Front showed that the
4th Ukrainian Front had been inserted between the 1st and 2nd Fronts. In addition, the new Front HQ could be located by
Direction finding (). :The Central Authority Networks were subdivided into; :* Central Authority of the Security Troops. :* Central Authority of the Frontier Troops. :* Central Authority of the Security Troops (Divisions and Brigades of the Back Area). :* Central Authority of the Railway Troops. • Networks of Formation. This covered the Front staffs with the
Grenztruppen Regiments. Communication between commanders of the security troops at the
front and their regiments, was monitored. D/Fing of the Regiments revealed the approximate extent of the Front. D/F of the Chief of Security Troops with the Front HQ. Finding the High Command indicated the approximate location of the Front HQ. Besides, when a forward echelon of the Chief of Security with the Front HQ appeared, it was possible to deduce that the Front HQ would move. This, on the basis of NKVD traffic alone. • Regiments to Battalions. Most of the messages could be read, they mentioned Army units by name, e.g.; :* Before the attack from the NEISSE sector (1st Ukrainian Front), a 2nd Army and 3rd and 4th Field Replacement Rifle Regiments were mentioned by III Bn.of the 83rd
Grenztruppen Regiment, operating in the
Görlitz area. The conclusion drawn was that the
2nd Polish Army had been newly brought up to the NEISSE sector. :* Lines of advance and boundaries of battalions were mentioned. Battalion boundaries often proved to be Army boundaries. Traffic of rear NKVD troops and of the Signals Regiment were of no interest, and were not covered by KONA 1. A constant watch was kept on the 1st network,
Central Authority of the Security Troops. It consisted of the communications between the Central NKVD Authority in Moscow and the commanders of those troops in Army Groups, directing staffs North and South, the less interesting GHQ Signals Regiments. These messages were not readable. Apart from the characteristics already noted they were also to be recognised by their use of Call-signs which were made up from a square. The Russian wireless station numbers were: The method of working was controlled () circle working. The Kreise were known by the number of the controlling station. They were: • Kreis VIII: Station 8 with 21, 23, 23 Night Frequency on 3170/3380
kcs • Kreis IX: Station 9 with 20,21,25,28 Night Freq. on 3850/4230 kcs. • Kreis XI: Station 11 with 26,29,87 Night Freq. on 4580/4940 Day Freq. 7400/8670 kcs. • Kreis XII: Station 12 with 27,84 Night Freq. on 4580/4940 kcs. • Kreiss LXXXI: Station 81 with 20,21,22,23,24,25,28,49,50,55,57,59,60,61,63,65 Night Freq. 2850/2975 kcs. Day Freq. 4080/4480 kcs. • Kreiss LXXXV: Station 85 with 26,27,29,51,52,53,54,56,58,62,64,80.84,87 Night Freq. 2700/3050 kcs. Day Freq. 3950/4400 kcs.
Call signs were mostly pronounceable. The Front Staff networks usually changed their call-signs daily, the Regiment networks at regular intervals of several days, in many cases of weeks. Frequencies used lay mainly between 2400 & 3600 kcs. It should be mentioned that 5-figure NKVD messages had been picked up in nets identified as belonging to the Army or the Air Force, such messages contained
SMERSH in the preamble. Such messages were originated by units of the
Red Army counter-intelligence units. Lt. Schubert spoke of two 5-figure codes, the
SMERSH organisation code, used with an individual subtractor, and the Railways Troop codes, that was a 4-figure code, enciphered with substitutions tables, the 5th digit representing the quadrant on the page in which the group appeared. Schubert was more familiar with the 4-figure NKVD codes, which were apparently more exploitable than the 5-F codes. The Security Troops Codes were of two types. One was used forward of regiment, when a regiment was used in approximately an Army Group Sector, i.e., it is a cipher, therefore, used forward of Army Group. This particular code ran for comparatively long periods, at least 1.5 years. The code, a
Substitution cipher, was alphabetical and contained 100 pages, with 25 or 50 groups per page, and enciphered by means of an enciphered
indicator, which provided for page and position substitution, the result of which was again enciphered by a substitution table. These substitution tables were also valid for a longer time and varied with the network. The additives on the page did not change, only the substitution table. The second type, a code used rearwards of regiment, was enciphered by a figure subtractor originally from tables (up to September 1944). The same tables could appear on different networks, and since the subtractor was used very frequently,
it was not uncommon for 20 messages to have the same subtractor In October 1945, 2 subtractors were used, taken from different tables, and the indicators for the 2nd subtractor were enciphered with the first one. In spite of the potential difficulties involved in this method on encipherment, the system was solved by 15 February 1945, though mainly because of bad Russian usage of the system. The Frontier Troops Code was exactly like the Security Code rearwards of regiment, with a different basic book, at least in traffic out of
Saint Petersburg (Leningrad) Lt. Schubert made a statement on the Railways and Convoy Troops code quotes above. The 4-F NKVD codes exploited by
KONA 1 were used by front line units, i.e. regiments and battalions, mainly employed by the Soviet Military Police. From the period of 1933 to 1942, code
R4ZC4 Russian 4-Figure, Code 4/ was in use. It consisted of a 100-page alphabetic book each with a hundred lines. It was broken by Germany in 1940. Only one code book, known to Germany as R4ZC 1800 and to the Russians as
KODOWA TABLICA ZERNO was used by the NKVD from October 1943 until the end of 1944.
NKVD R4 Z C 1800 Code The
R4 Z C 1800 code, had since October 1943 had 50 pages, each consisting of 50 lines and totalling 2,500 clear groups, alphabetically arranged. Each page is shifted by means of Cipher square () and the
bigrams (ab and cd) are then replaced by others according to a substitution table. The recipher consists of two elements:
Shifting by means of Chiffranten enciphering: Any figure, chosen from Row A (sample code in Annexe 1), will indicate on any page, in Row B below, the figure must be added to that standing beside the clear group (and subtracted when deciphering).The
chiffrant consisted of the figures 0 to 9 in random order printed at the top of the page, a different order being used for each page. The third figure of the 4-figure indicator group gave the number of the chiffrant to be used, i.e., if 5 was the 3rd figure of the indicator group, the fifth figure of the
chiffrant reading from left to right would be the one used for each page. This number was then added to the numbers of the lines on the page before they were reciphered by means of the bigram table, i.e. if the number was 6 then line 00 would become 06 and line 24 would become line 05.
Substitution of bigrams: For the first and second halves of the 4-figure group (elements
ab and
cd) there are substitution tables, each of which contains 100 bigrams (each bigram from 01 to 50 occurs twice). There are 10 such tables (numbered for the most part from 0 to 9). Two different bigram substitution tables were used for reciphering the big ram representing the page and that representing the line. 10 such tables were in use concurrently for the recipher of both page and line. Each table was 10 x 10 so that each bigram could be reciphered in two different ways and each table was designated by one of the figures from 0 to 9. The figure designating the table used for the line was put in the second place of the indicator group and that designating the table used for the page in the fourth place. The first place was a dummy. A series of substitution tables was current for a period of from 2 to 6 months. The reciphering is indicated by a 4-figure group (indicator) which appears at a definite position in the message and contains; • Number of the clear-group table (indicated with two digits). • Chiffrant. • Number of the pagination table. :::Example: :::Indicator 0151 :::Chigruppe 2406 :::Decipherment (according to the tables below) Element ab (=24) yields 20 according to the pagination table. The clear-group therefore is on page 20. Element cd yields 02 according to the clear-group table. From this the chiffrant is subtracted, in this case 1 (the number of page 20 under the fixed figure 5). Thus the clear group 2001, which represents
Komandirowatj is derived. The indicator group was inserted
en clair in one of the first ten groups of the message according to instructions. The penultimate group consisted of the date and length of the message and the last group the Chi number. R4ZC4 was used from 1933, solved in 1940 and in use until 1942. There were 100 pages each with 100 clear groups arranged alphabetically. Recipherment was by means of 31 substitution tables.
NKVD Other codes Though the emphasis by the POW's being interrogated by TICOM was on 4-figure NKVD codes, it should be pointed out that a great variety of encipherments were worked on, though not all of them read. Corporal Karrenberg, gave the following breakdown of nets and types of codes used: • Administration networks: usually sent 5-figure messages and less often, 4-figure and 5-letter messages. • Frontier networks: used 4 and 5-figure codes: on regimental networks, also 2-figure systems. •
Black Sea Fleet: used 3-figure and 5-figure codes networks from divisional level downward: used 2-figure and 3-figure, plain language and figure messages with plain language. Corporal Exeter of NAA 11 stated that he worked on 2-figure NKVD codes. Traffic of an NKVD net on the
White Sea Front, controlled from the NKVD HQ at
Archangel was worked on with some success.
NKVD Conclusions In general, certain elements seemed to be constant in all NKVD codes: • Simply that NKVD messages were enciphered on the same system. • NKVD messages were always arranged alphabetically. • NKVD messages whether 5-letter or 4-figure or 5-figure, had the date in the penultimate group.
Agent codes Agents' Codes included codes used by agents, guerillas and scouts () All systems were used from
substitutions, double
transpositions,
grilles and subtractors to one-time pads (tape based). Solving agents' traffic was always done centrally in Berlin. The solving of agent, guerilla and scout traffic was the responsibility of Referat IIIc of Group IV. Solving the cipher depended mainly on captured material and there was generally enough material to accomplish solutions. Some agent traffic was one-time tape and therefore unbreakable. By and large the group did not place much value on agent traffic and neglected it. It should be pointed out that Russian agent systems were not handled exclusively by OKH/GdNA/IV/3c, but were also handled by: • Section III of the Armed Forces Signal Communication Group,
Radio Defense Corps, or more accurately the AgWBNV, Radio Communication Group 3 (), (AgWNV/FU III) who were responsible for locating, eliminating, or neutralizing all enemy agents' radio activities. For more difficult cryptanalysis, traffic was turned over to OKH/Chi and worked on in
Referat Vauck, named and managed by
Wilhelm Vauck. • Some agents traffic was picked up independently of FU III, by the regular police, the
Ordnungspolizei, known colloquially as the Orpo, who sent their traffic to Kurt Sauerbier at the
Research Office of the Reich Air Ministry (FA), Principle Department IV, Referat 9c. There existed an entirely independent relationship between the Orpo and the Sauerbier and Sauerbier senior officer. Sauerbier's direct senior officer in the FA was Specialist Wenzer, who was an expert in Agents' ciphers and was sent from FU III to the FA to assist Lt. Schubert, of the OKH/GDNA, to assist on Polish Resistance Movement Traffic, in January 1944. There was, consequently, a certain duplication of effort and a certain amount of confusion as a result of this arrangement. There were also conflicting opinions of the opinions of the successes achieved. Schubert reported that
Russian agents' systems were tried by Dr Vauck, who stated they could not be solved. Later, he said that they were digit substitutions and P/L enciphered with a one-time running key derived from a book. Schubert thought that in the middle of 1944, e.g. the Soviets had about 3,000 agents in the field, and it was impossible to pick out one system and say that it was used in one area. Moreover, Schubert pointed out that he himself had worked on only Partisan and Kundschafter traffic and knew of other systems only indirectly.
Agent codes technical description For the W/T traffic of the Soviet Partisans and the spies there were used: • Double Transposition • Subtractor Reciphers • Occasional simple substitution systems The subtractor systems consist of 3 Cipher elements: • The basic cipher i.e. the substitution of the plain text by a substitution system • The recipher with the figure subtractor • The reciphering of the
indicators The Basic Cipher • On certain links, a 3 or 4-figure code was used as basic cipher. • In General, however, simple substitution systems were used, which substituted a 1 or 2-figure number for each letter, and were such that seven letters, which usually formed a key-word, were substituted by single figures, and the other three figures were used as tens for the other letters. The following is an example: Special points of this substitution system are as follows; • When plain text is enciphered by this system, in the cipher text 2 mostly, or all three, tens-figures occur with particular frequency. The frequencies of the figures vary. • A letter is not substituted by two like letters. For example., the squares 22,44,66 are empty. This in normal text, no figure can come three times in succession. • Numbers are recipher by repeating each figure three times: e.g. 1945=111999444555. :There are two ways of using this system: • The basic cipher is the same for all messages. • The basic cipher is variable, i.e. it changes from message to message, with the indicator. The indicator is constructed of five different figures. • These are written over the letters of the key and completed with the remaining five figures. The tens figures are taken from empty squares. The following is an example, with an Indicator of 37245 • There are occasional simple substitution systems which substitute each letter by a 2-figure number.
The Composition of the Subtractor The figure subtractors used are of three different kinds: :* They are printed on T/P rolls :* They are taken from tables :* They are built up from an indicator • The subtractors printed on T/P rolls were most frequently used and increasingly replaced the other systems, and had the soviet name of
Blocknot rulon. The instructions for use stated that each reciphering strip was to be destroyed when it had been used once. There were different rolls for
in and
out messages. Thus an unbreakable individual reciphering is achieved. A subtractor group from the roll, which is not used for reciphering and is sent in the clear, gives the starting point on the strip of the roll used. :Research into regularities in captured subtractor rules showed that: :* The rolls are made up on several machines, clearly show by the use of several type-faces. :* Longish repeats do not occur in a roll. ::The rolls have 5-figure numbers, which are given in clear as indicators in Partisan traffic, but not in Spy-traffic. • The use of subtractor tables comprises in the main, 4 different systems: :*There is a simple reciphering with a table of 100-300 groups. The starting point is given by a indicator which gives line and column of the first recipher groups: e.g. 11511=line 11, Column 5. Systems of his kind come especially in the traffic of
NKGB Saint Petersburg (Leningrad), and were solved. :*There are 100 recipher pages. The first two figures of a 5-figure indicator, made up of five different figures, gives the page used. At the same time the basic cipher is changed by the indicator. This system too occurred especially in the traffic of the
NKGB Saint Petersburg (Leningrad). :* There is a double recipher. For this is a table of 30 lines of 10 groups. The first recipher is taken from the first 18 lines, the second from lines 19-30. The starting point of the first recipher is given by an indicators, as bullet point 1, the second recipher always begins with the first group of the 19th line. This was also solved. :* There were besides other isolated systems using independent double recipherment. The system is shown by example. Two groups are table from a simple subtractor table, at a starting point, given by the indicator. e.g. :: :The following groups in the subtractor table are next changed. This gives the recipher, but of course, has not been solved. :Note: The marking off into blocks of 1,2 and 3 figures was apparently done to transpose the figures in these blocks according to the key provided by the top line. • Figure subtractors arithmetically constructed. The subtractors are built up from a 5-figure indicator, which contains five different figures. The methods of building them are demonstrated by examples. :* A) Simple addition in columns. The most usual type of this subtractor construction are the following: There are two substitution series, e.g.: :: :The indicator, e.g. 27345 is written down then substituted according to a) and the result written underneath. Then the indicator is substituted according to b) and the answer written at the side of the right: :27345 69820 :92183 :The groups on the left are now added, and the result written under on the right. Then the two groups on the right are added, and the result put below on the left, and so on: :27345 69820 :92183 19428 :78248 60321 :79749 47987 :07208 ----- etc. :The groups are used as subtractor, starting from the second or third line. In some systems, the right hand column is pushed down a line, which was solved. In some cases, the substitution series b) is derived from a). by pushing it along one or more places. The case also occurred of their being only one substitution series, with addition in one column only. • Simple cross addition. Starting with the indicator, each pair of adjacent figures is added, and the result written alongside. The following is an example: :Indicator 27345 :Subtractor 27345 90794 97633 63969 :This system occurred only as an emergency cipher, that was solved. There is a variation, in which one skips a figure: :27345 51896 30426 :or one may add in normal fashion three times and then skips a figure twice: :27345 90735 97022 ..... :These last systems were in use in conjunction with a changing basic cipher. This was solvable. • Addition in columns with a key phrase. The basic key varies by the indicator. A key group or phrase is enciphered in accordance with the basic key so that five groups of five are produced. The indicator is written under the first of these groups, the total of these two groups is put under the adjacent one, and so on: : :The subtractor is obtained either by adding these two rows and every subsequent row being the total of the last two rows or by containing the addition. This was solved. • Cross addition with key phrase. By enciphering a key phrase with the fixed basic key, you have a 5 -- 5-figure groups. The indicator is written under the first of these groups; as the next group you take the missing five numbers in ascending order and by cross addition, the five groups are filled in. The first row of the subtractor is produced by adding both rows. For example: : :By cross addition of the groups of the first subtractor row, four groups are formed for each original group and these are entered underneath. You then get a block of 25 groups. : :From the first group of the block, a key is made up by indicating the figures according to numerical orders by the numbers 1-5. :23451 :66793 :In accordance with this key, the columns are read out from the columns of the first subtractor block, commencing with the right column. From this, the second subtractor block is created: : :The third subtractor block comes correspondingly from the second by re-arranging the second group of the first subtractor row. This was solved. In a variation of this system, the first subtractor block is composed differently, the first subtractor row is exactly the same. The cross addition from the indicator is, however, extended to six groups. The sixth group is put under the first subtractor group. After the formation of the subtractor proceeds as described under c). This was solved. In this last type variable basic keys are also used. • Substituted cross addition. The indicator is substituted by a table. Five groups are formed by cross addition from the result. These were then converted into the same substitution table. From this, the first subtractor row is obtained. • Subtractor Boxes. First the figures 1-0 are written down and underneath the indicator and remaining figures. By adding every 2 rows, 9 further rows are formed. The 2-11th row are numbered 1–0. The following is an example. : :Then figure pairs are extracted from a particular row, e.g. row 3, and these indicate from which point each 5-figure group is to be read out of the box. This produces: :56=31420 :69=70169 :92=05050 etc. :This system appeared in conjunction with the variable basic key. A variation consists in a key phrase being enciphered on a basic key and from this, the first rows of the box are formed. This was not solved. Recognition of the subtractor system and its solution was achieved by the appearance of similar message endings and beginnings.
Indicators Indicators for the subtractor, based on the roll, were always inserted
plain. In the other systems there were several ways of putting in the indicator. Apart from a few exceptions, there were always two indicators. Possibilities regarding reciphering of these indicators were as follows: • Indicator groups plain • A certain 5-figure number is added to every indicator and the number is a constant • A certain group of the message is added to every indicator group. • Like bullet list 3, with the different that groups in the message are converted according to a substitution table • Like bullet list 4, where different substitution tables are used for the two groups of the message to be converted • Like bullet list 3, except that figures of the message groups are arranged in order and then added to the indicator group • Certain figures are extracted from several groups of the message, which are almost invariably the first five groups, for instance, the last figures of the first five groups and then added to the indicator group • Like bullet list 7, but substitution the adders in accordance with a substitution table. • Combinations of two of the above methods, doubled recipher of indicator groups, e.g. 2 with 3, 2 with 4, 2 with 5, twice 3. In general both indicators are reciphered on the same process, where one indicator is included at the beginning and one at the end of the message. The groups used for enciphering are also taken one each from the beginning and the end of the message. ==Miscellaneous cryptanalysis==