|
|
|
|
|
|
|
Enigma Family tree → Principle →
|
The rotor-based cipher machines
|
|
|
Invention of the Rotor Machine · 1915
|
|
|
|
The history of the Enigma starts around 1915, with the invention of the rotor machine. As usual in history, the rotor machine was
invented more or less simultaneously in different parts of the world.
Between 1917 and 1919 there were inventions from Edward Hebern in the US,
Arvid Damm in Sweden, Hugo Koch in The Netherlands
and Arthur Scherbius in Germany [1][2].
|
There is one development however, that pre-dates the others,
and that is the invention
of Theo van Hengel (1875-1939) and Rudolf Spengler (1875-1955), two
Dutch naval officers who produced working
rotor-based cipher machines
for the Dutch War Ministry (Ministerie van Oorlog) in 1915.
This novum was described in detail by Karl de Leeuw
in Cryptologia in January 2003 [2].
|
-
The name 'Enigma' was not yet used at that point.
-
Literally translated: test machine or prototype machine.
-
Apart from electricity, this patent describes the use of air, oil,
mechanics and light as the information transport medium.
It was important for the development of the
Štolba cipher machine in the mid-1930s.
|
|
In 1919, Scherbius started the development of a
rather large typewriter-style cipher machine,
which became known as
Die Handelsmaschine.
It was marketed from 1923 by
Scherbius & Ritter of Berlin-Wannsee (Germany),
and built by the company Gewerkschaft Securitas,
also of Berlin.
|
|
There were a lot of problems with the printing Enigma machines.
The first ones had reliability issues with the print wheel,
and the later model with the type bars. As a result, the
schreibende Enigma was not
suitable, reliable and affordable enough for the German Army (Reichswehr).
|
For this reason, Scherbius took his early
glow-lamp-based based Probemaschine and improved it.
The first new model was Enigma A,
and was introduced in 1924. It also became known as
Glühlampenmaschine (glow lamp machine).
The machine was available for about 1/8th of the price of the printing Enigma
and cost 'just' RM 1000 1 . It was housed in a wooden transit case
and resembles later Enigma models, except for the fact that the keys
and lamps are arranged in sequential order
(ABCDE...) rather than the common typewriter order (QWERTZ...).
|
|
|
Enigma A was succeeded by Enigma B (1924-1925) – of which various variants
were made – and eventually by Enigma C (1925).
The latter has 26 keys (A-Z) for the input, and 26 lamps
(A-Z) for the output. The text is scrambled by means of three cipher rotors
that protrude the top lid. Each cipher rotor has 26 contacts at either side.
Several variants of Enigma C were produced, such as the so-called
Funkschlüssel C (for the German Navy) and a Swedisch variant,
both with 28 keys.
➤ More about Enigma A
➤ More about Enigma B
➤ More about Enigma C
|
-
In 1924, the currency in Germany was the Reichsmark (RM).
|
|
Unlike the printing Enigma, the glowlamp machines had a reflector (UKW)
that made the machine reciprocal (symmetric). As a result, the settings
of the machine for encoding and decoding were identical, which greatly
reduced complexity. The UKW had two or four fixed positions.
The idea for the reflector came from Scherbius' colleague Willy Korn,
who would later lead the company.
|
In 1926, the design of the glow lamp Enigma was drastically improved.
A new chassis was developed and the standard (German) keyboard layout
(QWERTZ...) was introduced. Furthermore the reflector (UKW) could be set
to 26 different positions. It was mounted to the left of the three cipher
rotors, which is why this machine is sometimes thought to be a 4-rotor
Enigma.
The machine was known as Enigma D or Enigma Model A26.
Like its predecessor – Enigma C – it was housed in a wooden
transit case with a hinged lid, but it had several improvements.
|
|
|
The rotors could be accessed more easily — the top lid had a hinge
at the rear — there was an optional green filter for the lamp panel, and
it had a power selector that was located to the right of the cipher rotors.
Enigma D became the basis for most of the later Enigma designs.
Enigma D evolved into a line of commercial machines known as Enigma K,
with a range of variants. Note that the Type name 'Enigma D' was rarely used by
the manufacturer, as in 1926 model numbers were introduced. Enigma D was
known as A26 and its successor – Enigma K – was known as A27.
➤ More about Enigma D
➤ More about Enigma K
|
|
|
Cryptanalysis of Commercial Enigma · 1927
|
|
|
|
In late 1926 or early 1927, the British codebreaking establishment
GC&CS acquired a commercial Enigma D (A26), which was subsequently
analysed by codebreaker Hugh Foss [12]. In 1927
he wrote a detailed report [13], which included a method for breaking it,
provided that a piece of known plaintext (at the time known as a crib)
of sufficient length was available. With just 180 characters of known
plaintext, Foss could work out the wiring of the rightmost two rotors.
|
When the wiring was known, a crib of just 15 characters was
sufficient to recover the settings.
Foss' method was based on geometrics rather than algebra, but it showed
that the commercial Enigma could successfully be attacked. His method was
later improved by colleague
Dillwyn (Dilly) Knox, who developed two new
codebreaking methods called Rodding and Buttoning Up.
In 1937, during the Spanish Civil War, it enabled himself (Knox) and
fellow codebreaker William Bodsworth to break the Enigma K that was used by
the Spanish Navy and the Italian Navy.
|
|
|
Enigma K} was very similar to Enigma D, but had one important improvement:
The stepping notch — needed to advance the adjacent rotor — was attached to
the letter ring rather than to the rotor body. As the main commercial machine,
Enigma K was sold to a variety of national and foreign customers,
including the Swiss Army and the Swiss Foreign Ministry.
In addition, several wartime machines were derived from this model,
including the Japanese Army variant Enigma T (Tirpitz).
➤ More about Enigma K
|
|
In 1927, a series of new developments were started, all based on the
Enigma D design.
First of all there was the Commercial Enigma A27,
which later became known as the Enigma K.
There were several variants of this machine, such as the
Swiss Enigma K that was built for the Swiss Army.
|
All commercial Enigma machines had a simple rotor turnover mechanism
that is comparable to the odometer of a car. The rightmost rotor makes
a single step on each key press. After the rightmost rotor has completed
a full revolution, the middle rotor makes a single step and so on.
At the same time (1927) the development of a more advanced range of
machines was started. This range became known as
Zählwerk Enigma
(counter Enigma) or Zählwerksmaschine
(counter machine), probably because
it has a counter that shows the length of a message (key presses).
|
|
|
|
Furthermore, the Zählwerk Enigma
has a far more advanced rotor turnover
mechanism that is driven by cogwheels rather than by pawls and levers.
This allows the mechanism to be wound back as well,
which is useful for correcting mistakes. The Zählwerk Enigma
also introduces the concept of multiple turnover notches, which causes
more frequent (irregular) rotor stepping.
|
The three cipher rotors have 17, 15 and 11 turnover notches respectively,
each of which are relative primes of 26, which increases the period of the
machine (i.e. the number of steps before the sequence is repeated).
Furthermore, the reflector (UKW) is now part of the stepping mechanism
and is driven by the other rotors.
The first Zählwerk Enigma was Model A28.
It was introduced in 1928 and was built on the designs of Enigma D
and Enigma K.
A few years later, a design variant of the A28 was developed.
It was slightly smaller and had smaller cipher rotors.
|
|
|
|
Probably from the beginning of Scherbius' developments in 1917, the German
Army (Reichswehr, from 1935 known as Wehrmacht), had opted for a machine
that could be used in the field and that was secure enough to keep
messages secret for longer periods of time. Scherbius knew that the best
possible security would be obtained by increasing the number of rotors,
but came to the conclusion that this was not a viable solution due to
mechanical and practical contraints.
|
In 1926, development was started of a machine for the Reichswehr, known
as Enigma I. Based on the chassis of Enigma D/K, it had
3 rotors and a fixed UKW. Furthermore, it had a unique feature,
known as the Schaltbrett (switch board),
that significantly increased the strength of the cipher.
The first version of the Enigma I was ready in 1927 and featured
a single-ended Schaltbrett in a 2 x 13 circular arrangement.
After several redesigns of the Schaltbrett – by then
known as Steckerbrett (plug board) – they settled in
1928 for a much simplified double-ended plugboard.
|
|
|
|
It had been designed by the Reichswehr themselves, and is by far the
weakest variant of all tested plugboards. The final version of
Enigma I, with the double-ended plugboard, was put into service
on 1 June 1930. The plugboard was used exclusively on military
machines, and was not sold to commercial customers.
In 1932 the German Army claimed the exclusive rights to the machine,
after which all commercial and international sales had to be approved
by the German Army.
|
Meanwhile, in 1929, manufacturer ChiMaAG had successfully produced
a new and improved version of the ill-fated schreibende Enigma.
It was known as Enigma H or Model H29 and was designated
Enigma II by the Reichswehr. It was not compatible with
the glow lamp Enigma I.
The Enigma H29 was also sold to the Hungarian Army, but was never in
popular demand due to its high price. Apart from the Hungarian Army, the Germans
also kept selling Enigma machines to the Swiss and to the Dutch Navy.
It is known that the Dutch bought Enigma G
as late as 1938.
|
|
|
|
|
Numbers-only machine · 1930
|
|
|
|
A rather strange version of the Enigma, with a much smaller enclosure and
smaller rotors, was Enigma Z, or model Z30.
It was suitable for numbers only (0-9) and was intended for numeric messages
like weather reports and forecasts. Only a small quantity of this model was built.
|
The rotors of Enigma Z have only 10 contacts at either side. Furthermore,
there are 10 keys on the keyboard and 10 light bulbs on the lamp panel.
It is known that this machine was sold to Chile and Sweden, and that it was
offered to Spain and The Netherlands. After one or two production runs,
this model was discontinued.
It should be noted that initially, all glow lamp Enigma machines had sequential
Type Letters: A, B, C and D. In 1926, with the release of the Enigma D,
model numbers had been introduced. From then on, all machines – with the exception
of the military machines – were identified with model numbers like A26, A27
and A28. Around 1930, Type Letters were re-introduced, but this time they were
more meaningful than before.
|
|
|
|
Type Letters were used a a prefix for the model number, but also as a
prefix for the serial number. The first machines with this kind of
Type Letter were the Z30 (Enigma Z) and G31 (Enigma G). From 1936
onwards, the serial numbers of the A27 machines (Enigma K) were given
the prefix 'K'.
Likewise, Enigma T — manufacturered especially for the Japanese Army —
received the prefix 'T'.
Although the following is by no means certain, we believe the meaning of the
letters to be:
|
K Kommerziel (commercial) Z Ziffern (numbers) G Getriebe (gear) T Tirpitz
|
The first Enigma machines were developed by Arthur Scherbius
between 1917 and 1920,
but were built after the company had been dissolved into
Gewerkschaft Securitas and a few years later into
Chiffriermaschinen AG.
After Scherbius' untimely death in 1929, the company changed
hands, and in 1935, after the German Army had acquired the manufacturing
rights to the Enigma machine, the company's assets were taken over by
the newly established Heimsoeth und Rinke.
In the history of Scherbius and Enigma, the following events are
of importance [11]:
|
1911 Dipl.-Ing. E. Richard Ritter & Co (household appliances) 1915 Scherbius member of the telegraph troops 1917 War Ministry order for development of a cipher machine 1918 Arthur Scherbius, 1st patent (prototype) 1920 Scherbius & Ritter 1921 Gewerkschaft Securitas 1922 Securitas GmbH (workshop, Berlin, Germany) 1922 N.V. Ingenieursbureau Securitas (Amsterdam, Netherlands) 1923 Chiffriermaschinen Aktiengesellschaft 1929 Scherbius dies in accident 1935 Heimsoeth und Rinke
|
From 21 November 1921 to 1925, Scherbius was not in the management
of the above listed companies and did not own the majority of shares,
except for his part in Scherbius & Ritter [11].
➤ Enigma related patents
|
|
|
Other Enigma manufacturers
|
|
|
As many Enigma machines were needed for the German war effort,
other companies were contracted to build the machines under
licence. This also reduced the risk of supply problems should
any of the manufacturers be bombed by the Allies.
With Heimsoeth & Rinke in Berlin being the engineering
company and main contractor,
the machines were manufactured by Konski & Krüger in Berlin.
Later, the military machines were also manufactured by
Olympia in Erfurt,
Ertel-Werk in München and
Atlas-Werke in Bremen, all under licence of
Heimsoeth und Rinke.
➤ List of manufacturers
|
|
|
The Polish Breakthrough · 1933
|
|
|
|
Around 1930, the Polish Cipher Bureau, Biuro Szyfrów, was the first
to make an attempt to break the military Enigma. As one of the closest
neightbours of Germany, they were very much aware of the clear and present
danger of another war and started researching
commercial Enigma K (A27).
|
|
|
Marian Rejewski
|
|
Jerzy Rózycki
|
|
Henryk Zygalski
|
|
From the University of Poznan, three young brilliant mathematicians
were recruited: Marian Rejwski, Jerzy Rózycki and Henryk Zygalski.
They started working on the Enigma cipher with nothing more that a
handfull of intercepted messages and a description of a
commercial Enigma.
|
Rejewski was set to work on the problem in late 1932, and after a few
weeks he achieved his first breakthrough, when he deduced the secret internal
wiring of the Enigma. Together with his colleagues he started developing
various aids for the regular decryption of German Enigma traffic.
Zygalski developed the so-called Zygalski sheets that were used to
exploit the double-enciphered message indicator, 1
a weakness in the German procedures. Later a machine was developed to
exploit this weakness mechanically: the
Bomba kryptologiczna —
the cryptologic bomb.
|
|
|
With only three rotors available for the Enigma I (in 1933), there are 6
possible rotor orders. On the Bomba, six sets of Enigma rotors were driven
simultaneously by a cog wheel at the center. Around 100 intercepted messages
were needed to recover the rotor order and initial settings.
➤ More about the Bomba
|
-
The double-enciphered message indicators — that had enabled the Poles
to break a significant part of the Enigma traffic — dates back to the late
1920s. It was part of a proposal on how to use the commercial Enigma.
In early 1940, newly hired mathematicians at OKW/ln7 discovered the
weakness that it introduced. The procedure would eventually be
abandonned by the Germans on 1 May 1940, but by that time WWII had already
started and the British had become involved in codebreaking (see below) [10].
|
|
In 1933, the Polish Cipher Bureau got access to the Enigma operating
procedures that were used by the German Army. Hans Thilo Schmidt, a German
playboy working at the German Cipher Office, needed money and sold
infomation to the French secret service. The French, who gave Schmidt the
codename Asché, passed it on to the Poles,
who could then reconstruct the machine.
|
From 1933 onwards, the Poles intercepted and decrypted a significant
portion of the German radio traffic. In 1938 they noticed an increase
in the number of messages; an indication that the
Germans were probably preparing for war.
Until that point, the Germans had been using a common Grundstellung
(basic setting) for all traffic. On 15 September 1938 however,
this procedure was abandonned. Around the same time, two new rotors
(IV, V) were added to the existing three, which multiplied the
maximum number of possible settings by a factor of 10.
|
|
|
In the meantime, the Poles had built their own equivalent of the
Wehrmacht Enigma with a plugboard added towards the rear.
The wiring of the two additional rotors was soon recovered by Rejewski
and suitably wired rotors were added to the Polish Replica. With the
war imminent, the Poles started looking for ways to get their knowledge
out of the country before it was too late.
|
Dilly Knox
was one of the British codebreakers of Room 40 during
World War I.
Since 1925 he had been trying to break Enigma, and had his first
success on 4 April 1937 when he broke Franco's
Enigma K during the Spanish Civil War.
When Germany started using Steckered Enigma
for communication between Germany and Spain in 1938,
he mounted an attack on the military Enigma, but did not succeed,
as he was unable to work out the wiring of the entry disc (ETW).
In 1938, GC&CS started discussing Enigma with the
French cipher bureau – Deuxième Bureau – from whom they acquired the details
that the French had obtained from the German spy Asché.
The French also disclosed their contacts with the Poles.
In January 1939, at the first Polish-French-British meeting in Paris (France),
GC&CS was represented by Dilly Knox,
Hugh Foss
and Alastair Denniston.
Knox described the system of rodding that he had developed, but the Poles
had been instructed by their superiors not to disclose any vital information
at that meeting.
|
Dilly had clearly impressed the Poles and on 25-26 July 1939,
with the war imminent, a second meeting was arranged, this time in Poland at
a facility of the Polish Cipher Bureau in a forest near Pyry, south
of Warsaw (Poland). At this meeting, the Poles revealed their achievements.
Also at this meeting, the Poles gave a
replica machine
to both the French and the British.
|
|
|
Rejewski had used a different approach to Knox, as he used (mathematical)
permutation theory to solve the problem, whilst Knox applied linguistics.
Nevertheless, the two quickly established a good relationship during
the conference. Knox also learned that the Enigma entry disc was simply
wired in alphabetical order. Something that neither he nor
Alan Turing had ever considered.
|
|
The meeting in Pyry was attended by Dilly Knox
as codebreaker, Alastair Denniston
as head of GC&CS (and codebreaker), and Humphrey Sandwith
as head of the Admiralty's intercept and
direction-finding service.
On behalf of the French Deuxième Bureau, Captain Gustave Bertrand was present.
The Polish contribution would soon prove to be of vital importance
to the war effort.
|
Immediately after the meeting, the Polish Cipher Bureau
destroyed all their secret documents and equipment, whilst
the cryptanalysts escaped to France.
A few weeks later, on 14 August 1939, Bletchley Park
was established by the British.
Only two weeks later, on 1 September, Germany invaded Poland and
two days after that, on 3 September, the UK and France
declared war on Germany. World War II
had started, just five weeks
after the Poles had shared their secrets. In the early stage of
the war, the Poles continued to work on Enigma from the French
Cipher Bureau.
|
|
|
|
Bletchley Park is an estate in the small town of
Bletchley (Milton Keynes, UK), some 45 miles north of London, that
became home to the
Government Code and Cypher School (GC&CS)
— the British Cipher Bureau. The location
was choosen because it had direct railway connections to London,
Cambridge and Oxford, allowing scientists and army personnel to travel
unobtrusively.
|
The first people to arrive at Bletchley Park (BP)
were professional codebreakers, chess players, mathematicians and people with
organising skills. Among them:
Dillwyn (Dilly) Knox,
Gordon Welchman,
Alan Turing
and Stuart Milner-Barry.
Knox had already worked for the codebreaking unit Room 40 during
World War I, and helped with the decryption of
the famous Zimmermann Telegram which brought the US into the war.
Stuart Milner-Barry was a chess player and chess writer. Gordon Welchman
and Alan Turing were both mathematicians from Cambridge (UK).
|
|
|
Enigma messages were initially broken by hand, using simple pencil
and paper methods, and with additional tools like the so-called
Jeffrey Sheets — the British equivalent of the
Polish Zygalski Sheets. But as the volume of the traffic
increased, Turing had to look for automated solutions.
Based on the Polish Bomba
and the information that had been passed by
the Polish codebreakers shortly before the start of WWII,
Turing developed the Bombe.
Although the Polish method of exploiting
the German weakness of the double-encypered message indicator
could no longer be used, Turing developed a more generic method
based on Cribs
(pieces of guessed plain text).
➤ More about the Bombe
|
|
The British Prime Minister, Winston Churchill,
recognized
the value, impact and importance of the intelligence delivered
by Bletchley Park (BP) and its codebreaking apparatus.
He introduced a new level of secrecy that superseded all others:
TOP SECRET ULTRA — abbreviated ULTRA — and demanded that the
source of this ULTRA intelligence be kept secret at all times and at all costs.
|
At the first stages of the war (1940), the British codebreakers were
able to read the majority of radio messages from the German Air Force
(Luftwaffe) and a modest part of the Army traffic (Wehrmacht). The
Naval messages on the other hand, posed a real problem as their operating
procedures were much more complicated.
Furthermore, the German Navy (Kriegsmarine) used three additional rotors
(VI, VII and VII)
of which the wiring was initially unknown. These extra rotors were used
exclusively by the Navy and were not shared with the other forces.
|
|
|
|
In 1941, Turing achieved a breakthrough when he was working in isolation
in The Cottage at BP.
He discovered the wiring of the additonal rotors
and the naval message indicator procedure. Aided by the catch of a large
amount of codebooks from U-Boot U-110, captured on 9 May 1941,
Turing managed to find a way into the Naval Enigma M3
and to decrypt part of the naval traffic.
|
Apparently, the Kriegsmarine used a complex procedure that involved
several codebooks, short message books and substitution tables.
Messages and status reports were shortened by translating them
into a short letter combination.
The British even develop a unique system for fast
radio direction finding, known
as Huf-Duf (HFDF)
in order to obtain useful cribs for the Bombe.
Then, on 1 February 1942, disaster struck when the German Navy,
completely out of the blue, introduced a new type of Enigma machine,
which caused an immediate and full black-out at BP.
|
|
|
The new machine had an extra cipher rotor, inserted
between the leftmost rotor and the reflector. At the same time,
the indicator system was changed and new codebooks were introduced.
The new machine – known as the Enigma M4 –
was used almost exclusively by the U-boat division of the Kriegsmarine.
The Bombes, that were made for the 3-rotor Enigma, were not suitable for this.
➤ More about Enigma M4
|
|
|
The Battle of the Atlantic
|
|
|
|
During WWII, there was en enormous shortage of nearly everything in the
UK. Large convoys of supply ships, the so-called Liberty Ships, travelled
from the US to the UK, bringing people, food, ammunition and anything
else that was needed, to wartime Britain and to the Soviet Union [6].
|
Althoug the Liberty Ships were designed in the UK,
they were adapted by the US and were quick and cheap to build.
They sailed in convoys and were protected by military vessels.
Nevertheless they were an easy prey for the German U-boats that were
organised in the so-called Wolfpacks.
The 4-rotor Enigma M4
had a serious impact on The Battle of the Atlantic.
As it was no longer possible to read the messages to and from the U-boats,
it was impossible to determine the location of the Wolfpacks,
resulting enormous losses of ships, people, supplies and war cargo.
|
|
|
|
The black-out that started on 1 February 1942, lasted for nearly nine
months and cost no doubt numerous lives. Luckily however, the tide
changed on 30 October 1942, when new codebooks were captured from a sinking
U-boat. In the meantime, Turing had already worked out the new Naval procedures
and the wiring of the additional rotor. The codebooks completed the puzzle.
|
As the Bombe
was only suitable for attacking 3-rotor Enigma machines,
several solutions were developed. A 3-rotor Bombe, that contained the
equivalent of 36 Enigma machines, was modified into a 24-Enigma
4-rotor Bombe. Although the resulting machine was rather slow, it worked.
A better solution was to add external 4th-rotor attachments
to the existing 3-rotor Bombes. Such add-ons were developed and built by
the British Tabulating Company (BTM) as well as by the
General Post Office (GPO) at Dollis Hill.
Some solutions even involved valve-based technology.
|
|
|
|
Finally, several variants of a true 4-rotor Bombe were built.
Some of these featured an extra fast 4th rotor and an electronic
valve-based sensing circuit that was developed by
Tommy Flowers
at the GPO in Dollis Hill. By this time, the US had already entered
the war and after long discussions it was decided to share the knowledge
about the Bombe technology with the American Allies.
|
This decision — that allowed the Americans to develop their own
4-rotor Bombe — came just at the right moment. As the UK suffered
shortages of nearly all kinds of material, it became more and more
difficult to build reliable machines.
The Americans on the other hand, had sufficient supplies and
resources and were able to allocate funding and production capacity.
The US Bombe
was developed by Joe Desch, an engineer at the
National Cash Registers (NCR) in Dayton (Ohio).
Development started at the end of 1942 and by mid-1943 the first
US Bombe was introduced.
|
|
|
It appeared to be much faster the UK Bombes
and involved valve-based electronic circuits.
By the end of 1943, no less than 120 machines were installed, and
for the remainder of the war, the US took care of breaking the bulk of
4-rotor Enigma traffic (i.e. message to and from the U-boats), leaving
a modest part of it, plus the bulk of the 3-rotor traffic, to the
codebreakers in the UK.
➤ More about the US Bombe
|
|
The most common version of the Enigma machine that was broken by the
codebreakers at Bletchley Park (BP)
was Enigma I,
the machine that was used by the German Army and Air Force.
The Naval machines, M1, M2, M3 and M4,
were also broken on a regular basis. Nevertheless,
there were other Enigma models and variants that required the attention
of the BP codebreakers.
|
Some less important networks, sometimes used Commercial
Enigma machines. Such machines, generally the Enigma K,
were also used by other countries,
such as Italy, Spain and Switzerland.
One of the most difficult machines to be broken, appeared to be
the Enigma G. It was a variant of the commercial Enigma
that had a cogwheel driven mechanism and multiple turnover notches on
each rotor, causing irregular rotor stepping.
Such machines were used by the German Secret Service,
the Abwehr
(hence the nickname Abwehr Enigma) and could not be broken by the
Bombe.
|
|
|
The Abwehr networks yielded far less intercepts than the regular army
networks, making it very difficult to find messages in depth. 1
Furthermore, the Abwehr used different keys on each link, requiring
each radio link to be broken individually.
The Enigma G was attacked by a team led by
Dillwyn (Dilly) Knox,
the WWI Room 40 codebreaker who had already broken the Spanish Enigma.
|
During the First World War (WWI),
he had helped to break the
famous Zimmermann Telegram, which was responsible for bringing the US
into WWI.
Knox was among the first group of people
to arrive at Bletchley Park (BP) in
August 1939, where he started working in
The Cottage.
After breaking the Italian Naval Enigma in 1941, something that was
decisive in winning the Battle of Matapan (Greece), he and his group of
female codebreakers – known as Dilly's Girls – started working on the
Abwehr Enigma, and by the end of 1941 they had their first success.
|
|
|
After the first breakthrough in October 1941, a special unit was established
to work on the Abwehr decrypts. It became known as
Intelligence Services Knox (ISK)
and by the end of the war,
ISK had processed about 140,800 Abwehr messages [7].
Knox himself didn't live to see the results of his work.
Already diagnosed with lymph cancer at the start of the war,
he died in February 1943.
One of 'his girls' — top codebreaker
Mavis Lever (later: Batey)
who had broken the Italian Naval Enigma in 1941 —
wrote an affectionate biography of
Dilly Knox
in 2009 [8].
➤ More about Abwehr Enigma
|
-
Messages 'in depth' means that two (or more) messages had been
encrypted with the same key.
|
|
Before and during World War II,
Japan was arguably Germany's most
important ally. Germany mainly fought its war in Europe, North
Africa and Russia, whilst Japan took care of the southern hemisphere.
During the war, Japan had observers in all parts of the European war theatre.
|
For communication between the observers and their headquarters,
the Japanese used two manual cipher systems, known as Sumatra
and TOGO (later: Sumatra 2 and TOGO 2), but they preferred a mechanical
system like the Enigma.
Although they preferred the Military Enigma I
(with Steckerbrett),
the Germans didn't want to give away their most secure Enigma machine.
Instead it was agreed in 1942 to build a special version of the
commercial Enigma K
with a differently wired entry disc (ETW)
and five turnover notches on each of the eight rotors.
|
|
|
The new machine was designated Enigma T (Tirpitz) and
the Japanese ordered 800 of them.
The first batch was delivered in August 1943.
Due to material shortages however, the full order was never delivered.
Furthermore, the Japanese had their doubts about the security of the
machine and insisted on having the military Enigma I instead.
It was agreed that the remainder of the order would
consist of Enigma I
machines that were made backwards compatible with
Enigma T.
➤ More about Enigma T
|
|
In retrospect it may seem strange that the Germans kept using Enigma
for so long and that its security was never questioned. In reality
however, questions about the cryptographic strength of Enigma had
been raised several times, in particular by U-boat commander
Admiral Karl Dönitz.
|
On each occasion, the Army Intelligence Service, (Abwehr)
was asked to investigate any incidents. But the Abwehr,
who had been responsible for choosing the Enigma in the first
place, always concluded that it was imposible to break the machine.
After all, the British used it as well...
Nevertheless, Donitz kept having doubts and took his own measures.
Three additional cipher rotors (VI, VII and VIII) were introduced
in 1939 for exclusive use by 'his' Navy and in 1942, out of the
blue, he introduced the M4 Enigma.
But those were not the only security measures.
|
|
|
In 1943, a new reflector Umkehrwalze C, or UKW-C,
(with 4th rotor 'Gamma')
was introduced as an alternative to UKW-B/Beta, but it was only
available to a limited number of users.
Nevertheless it was used until the end of the war,
sometimes even mixed with UKW-B and 4th rotor 'Beta'.
|
In January 1944, a rewirable reflector,
UKW-D
or Dora, was introduced. It could be fitted in place
of the existing UKW-B and there even was a special Naval version of it.
Codebooks were updated to include the UKW-D wiring, which was
changed every 10 days. Nevertheless UKW-D
saw limited use as it was difficult in operation
and could not be distributed effectively in 1944.
The German Air Force, the Luftwaffe, took their own measures and
developed a device – known as the Enigma Uhr – to alter the wiring
of the Steckerbrett (plugboard) for each new message.
|
|
|
|
The Uhr
was a small wooden device that could be attached to
the right side of an Enigma machine and had 20 wires that
were connected to the Steckerbrett instead of the
normal patch cables. A large wooden knob on top of the device
could be set to any of 40 positions, marked 00 - 39.
|
The Uhr was even combined with UKW-D
on the so-called Red key,
making it a real challenge for the codebreakers at
Bletchley Park.
Had the Uhr been used correctly, it might even have defeated them,
but due to operator mistakes it was broken within a few days
after its introduction.
The most dangerous Enigma improvement however,
was the so-called Lückenfüllerwalze
(gap-filling rotor). It looked and worked like a regular
Enigma cipher rotor, but had 26 configurable notches that could easily be
altered in the field. This way it introduced irregular rotor stepping.
|
|
|
|
Once the war was over, the fact that the the Enigma had been broken
was kept secret for many years. Apart from a few exceptions, people went
on with their lives and most of the Bombes
were dismantled. The captured Enigma machines ended up in the vaults of
GC&CS (later: GCHQ) and the SIS (now: NSA),
or were used by other countries in the believe that
they could not be broken.
|
In countries like Norway, Germany and Austria, the
Enigma-I was used
for many years after the war, until they were replaced by newer
and better equipment. Rumour has it that Enigma was also used in
some African countries.
There are no reports about the use of Enigma by the
Russians, although it is certain that several machines were captured.
For a long time, historians assumed that the Russians had no knowlege
about the Allied codebreaking successes during WWII,
but it has since become clear that they were well informed through espionage.
|
|
|
|
In 1956, the Russians introduced the first version of a sophisticated
rotor-based cipher machine: the M-125 —
codenamed FIALKA. The machine has 10 rotors
and features irregular stepping, whilst alternate rotors move in opposite
directions. More importantly, it has fixes for virtually all known
design flaws of the Enigma, including the fact that a letter cannot
be enciphered as itself.
|
Furthermore, the Steckerbrett is replaced by a card reader and
the machine can operate directly on teleprinter signals,
allowing the use of letters and numbers.
It has a built-in tape puncher and reader,
and prints its output directly onto a paper strip.
Available in several variants, it was used by the USSR
and all Warsaw Pact countries.
Shortly after the end of WWII, the Americans started the
development of a new rotor-based cipher machine that would replace
the wartime SIGABA. It became known as
KL-7,
and also by its key-procedure
names ADONIS and POLLUX.
|
|
|
|
KL-7 was used by all US Forces
(Navy, Army and Air Force), as well as by government agencies
like the FBI, CIA and the White House itself.
It also became the main cipher machine of the newly established
NATO in post-war Europe.
From the 1960s onwards, rotor-based cipher machines were
gradually succeeded by electronic alternatives, such as
KW-7, KL-51 (RACE) and
Aroflex.
|
Fialka and KL-7 were not the only
rotor-based cipher machines
that were based on Enigma. The Swiss Army had been using a
variant of Enigma K since 1938, but knew it was unsafe and
probably suspected the Germans to read their traffic.
The Swiss therefore decided to develop their own machine, which had
a similar appearance but was mechanically more advanced. They called it
'Neue Maschine' (new machine) — abbreviated NeMa.
Development started in 1941, but the new machine wasn't ready when the war
ended. Nema was eventually introduced in 1946.
|
|
|
By far the most sophisticated rotor-based cipher machine however,
was developed by Boris Hagelin in Switzerland,
the man who had made a fortune during WWII by selling his
M-209 pin-wheel cipher machine to the Americans.
For many years after the war, he played with the thought to develop a
'Super Enigma', that would surpass any other rotor machine in the world.
Development of the machine started in 1954, but was delayed several times due to
mechanical difficulties. As a result, the name of the machine also changed
a number of times (e.g. HX-57).
|
|
|
|
In 1963, the machine – by now known as HX-63 – was released,
but hit the market too late. Electronic cipher machines based on shift-registers
were rapidly taking over the markt.
|
Based on many years of research by Frode Weierud, we've been able
to put together the most accurate family tree of Enigma machines to date.
It shows the relationship between the various models and variants,
and provides a lot of additional information.
Please note that the tree is based on ongoing research and is therefore
subject to changes.
➤ More information
|
|
|
The history of the Enigma machine is extremely complex.
There were many different models and variations,
and they were used by many different customers.
During the war, a mixture of military and commercial
Enigma machines were used by different branches of the
war apparatus.
Based on the above research, we've created a timeline
of events, patents, enigma models, accessories and peripherals.
➤ More information
|
|
|
|
|
Other WWII German Cipher Machines
|
|
|
|
The Enigma wasn't the only rotor-based cipher machine
that was used during World War II. The UK used the so-called
Typex cipher machine
for all high-grade traffic. Typex — basically a copy
of the German Enigma — had five cipher rotors, three of which moved
during encipherment.
In fact, the British codebreaking centre Bletchley Park
had huts full of Typex machines that were converted into Enigma analogues,
to decrypt German messages once the key had been broken.
|
As far as we know, Typex
was never broken by the Germans during the war,
despite the fact that the Germans had captured some machines.
On the contrary: the fact that the British used a similar machine,
confirmed the German believe that the
Enigma was indeed unbreakable.
The Americans also used rotor machines for some of their
radio traffic. For low-grade tactical messages they used the
Hagelin M-209, which they knew could be
broken by the Germans.
For high-grade traffic however, they used the
15-rotor SIGABA, shown
in the image on the right.
|
|
|
A combined development of US top cryptographers William Friedman,
Frank Rowlett (US Army) and Laurence Safford (US Navy), SIGABA is based
on the same rotor principle as Enigma, but is improved in many ways.
In addition, it prints its output directly onto a gummed paper strip.
As far as we know, SIGABA was never broken by the Germans during the war.
In the latter part of the war, around November 1943, the need arose for the
Americans and the British to securely exchange cipher messages.
As they couldn't agree on which machine was the better one,
Typex
or SIGABA,
it was decided to define a common standard and modify both
machines to comply with that standard. The common machine became known
as Combined Cipher Machine (CCM).
|
|
|
Development and production
|
|
|
The timeline below shows when the various Enigma models were developed and
produced. The history starts with the Dutch invention of the
rotor machine in 1915, followed by several inventions of
similar machines in 1917. In November 1917, German engineer Arthur Scherbius
started the development of his contribution to history, which would later
become known as the Enigma.
From his early protype – the
Probemaschine (prototype) – he developed
a line of Printing Enigma Machines (1923) and a lines of cheaper and
smaller Glowlamp Enigma Machines (1924),
the latter of which appeared to be the
more successful one. Initially the models A, B, C and D
followed in quick succession, but that changed in 1927, when several new
models were developed.
The German Army and Air Force adopted Enigma in 1927, but the development of
what would become known as 'Military Enigma I' took until 1930. Until
that time, various designs of the Steckerbrett (plugboard)
were tried. Also in 1927, the development of a more advanced machine with
irregular rotor stepping was started. It became known as the
Zählwerksmaschine (counter machine).
The initial Enigma A28 eventually evolved into the smaller
Enigma G (G31), often referred to as the Abwehr Enigma,
as it was used by the German intelligence service,
the Abwehr.
The German Navy (Kriegsmarine) was already familiar with the Enigma machine
since 1926, when they purchased Funkschlüssel C
— a special version of commercial Enigma C that had been adapted for
the Kriegsmarine. It took however until 1934 to adopt the Military Enigma
in a slightly modified (but compatible) form. The first Naval machine from
this generation was the Enigma M1, followed by the functionally identical
M2 and M3.
Eventually, in late 1941, a four-rotor machine was developed
exclusively for the Kriegsmarine. It was designated M4 and was
backward compatible with Enigma I and
Enigma M1, M2 and M3. After this date, no new M3 units were made.
In the diagram above the rightmost column shows several special models,
such as the numbers-only Enigma Z30, the Railway Enigma
and Enigma T. The Railway Enigma was a variant of the
commercial Enigma K that was made for the Reichsbahn.
Enigma T was made especially for Japan and was also based on
Enigma K, albeit with modified wiring and multiple notches
on the rotors.
|
- Wikipedia, Rotor machine
Retrieved January 2014.
- Karl de Leeuw, The Dutch invention of the Rotor Machine, 1915-1923
Cryptologia, January 2003, Volume XXVII, Number 1, pp. 73-94.
Author's copy.
- Wikipedia, Enigma machine
Retrieved January 2014.
- Hugo Alexander Koch, Patent NL10700
7 October 1919. Transferred to Securitas on 5 May 1922. 1
- Dr.-Ing. Arthur Scherbius. Enigma Chiffriermaschine
Elektrotechnische Zeitschrift. 1923. Heft 47/48. p. 1035-1036.
- Wikipedia, Battle of the Atlantic
Retrieved January 2014.
- Wikipedia, Dilly Knox
Retrieved January 2014.
- Mavis Batey, Dilly, The Man Who Broke Enigmas
2009. Hard cover, ISBN 978-1-906447-01-4.
- Kruh and Deavours, The Commercial Enigma: Beginnings of Machine Cryptography
Cryptologia, Volume XXVI, Number 1, January 2002.
- Frode Weierud, Personal correspondence
August 2017.
- Claus Taaks, Scherbius and Enigma History
Personal correspondence, April 2023.
- Hugh Foss, Reminicences on Enigma
1949. Published in Chapter 3 of the book Action This Day.
Michael Smith & Ralph Erskine, 2001. ISBN 978-0-593-06357-6.
- Hugh Foss, The Reciprocal Enigma 1
TNA, HW25/14. Undated, but probably 1927/28.
|
 |
|
-
The rights to patent NL10700 were transferred to Naamloze Vennootschap
Securitas in Amsterdam (Netherlands) on 5 May 1922 and then to
Chiffriermaschinen AG in Germany on 28 January 1927.
|
|
|
|
Any links shown in red are currently unavailable.
If you like the information on this website, why not make a donation?
© Crypto Museum. Created: Wednesday 14 March 2012. Last changed: Monday, 11 November 2024 - 12:28 CET.
|
|
|
|
|
