The Sigaba cipher machine

Home Page
The Sigaba Home Page

Introduction

The Sigaba is a revolutionary American electromechanical cipher machine. It was put into service in 1941. It will remain in service until the end of the 1950s. The Sigaba was replaced by the KL-7 cipher machine.

This machine is one of the most famous of WWII, like the Enigma and the M-209. It was the machine of the time which offered the highest level of security. In fact, neither the Japanese nor the Germans have ever deciphered a single message encrypted by this machine (see "Security of Our High-Grade Cryptographic Systems" in the Web links section). Even today, with the help of super computers, a message encrypted by Sigaba is indecipherable if we respect the security procedures in force during WWII. This is not the case with Enigma: a cryptogram of a few dozen characters can be deciphered without any clue. (see the article about Enigma by Olaf Ostwald and Frode Weierud in Web links section)

In all, nearly 10,000 machines were built.

Fig. 1: Sigaba at Crypto Museum 2016 exhibition

Note: If the Sigaba was withdrawn from service at the end of the 1950s, it was not because its Security level was no longer high enough, but because the machine was too slow and also because it was not practical. In fact, it only allowed “offline” encryption. That is to say, it generated a printed strip corresponding to the encrypted text which then had to be transmitted in Morse code or using a teleprinter. The German T-52e encrypting teleprinter (contemporary of the Sigaba) made it possible not only to encrypt but also to transmit the generated ciphertext.

Names and models

The name SIGABA is the machine's code name for the US Army. This machine was used by both the US Navy and the US Army and later by the US Air Force. Its name was different depending on the US armed forces:

  • Navy: ECM Mark II
  • Army: M-134-C
  • AFSA: (the ancestor of the NSA which includes the Air Force in addition to the Navy and the Army): ASAM-1 (ASAM: Armed Forces Security Agency Machine).
  • NSA: KL-29 (formely CSP-2900)

The Sigaba existed in two models, which are known in the Navy under the following names:

  • CSP-888/889 put into service in 1941.
  • CSP-2900 put into service in 1945.

History, evolution

The genesis of Sigaba occupies almost the entire period between the two wars and is closely linked to the history of the cryptological services of the American Army and Navy.

The First World War demonstrated the inadequacy of the use of codes to encrypt messages. Indeed, using the codes took too long, generated too many errors and inaccuracies and finally updating and deploying them was a real headache. The main cryptologists of the time wanted to replace them with the emerging cipher machines.

Hebern, from 1921, presented to the American Navy a cipher machine equipped with a rotor. The rotor concept is revolutionary and was invented by Hebern and other inventors almost simultaneously (see Rotor). The Navy is very interested in this machine and pushes Hebern to improve it.

In 1924, the Navy was ready to adopt a version equipped with 5-rotors manufactured by Hebern. Before adopting it, the Navy asked W. F. Friedman, the famous cryptologist of the US Army, to solve a challenge composed of 10 cryptograms (see challenge). Friedman wins the challenge. As a result, the Navy abandoned its idea of equipping itself with Hebern's 5-rotor machines.

Then, in parallel, the Army and Navy will separately try to develop a cipher machine offering a high level of security.

After winning the Navy challenge, Friedman is convinced that a rotor cipher machine is the future of cryptography. In addition, he understood that the main flaw of Hebern's machine is the regular advancement of the rotors. He invented a machine (the M-134-A) equipped with 5 rotors whose advancement is controlled by a telex tape (see History of the M-134-C in the reference section). Unfortunately this machine, its predecessors and successors will take years to starting to be built. Indeed, the army had a very low research budget following the great economic crisis of 1929. In 1935, Friedman's assistant, Frank Rowlett, made a huge discovery: he proposed to Friedman to replace the Telex tape with rotors to direct the advancement of the encryption rotors. But, due to lack of credit, this innovation was not put into practice.

After the challenge of 1924, the Navy, which had no money problem, persisted in wanting to develop a rotor cipher machine. On the other hand, even if she continues to work with Hebern, she uses her own ideas to improve the advancement of the rotors to make it more random. She orders machines from Hebern and then transforms them under the direction of Laurence Safford (see the History of Mark II ECM in Web links section). In 1934, the Navy was almost sure that she had succeeded in building a very secure machine. She offers a new challenge to Friedman which the latter (helped by his team) wins (see Second Challenge). The Navy still puts the machine (called ECM Mark I) into service but using the most secure modes not broken by Friedman. She continued her theoretical studies and in 1935 asked the Army if it proposed a new approach to make the advancement of the rotors less predictable. With the agreement of his superiors, Rowlett revealed to the Navy the principles that he and Frideman had invented. Without showing ostensibly interest in the Army's ideas, the Navy developed a new machine which would become the Sigaba.

In 1938, the Navy unveiled this machine to the Army (see "History of Magic" in the Reference section). Friedman and Rowlett note the incorporation of their ideas into this machine. The Navy and the Army decide, after agreeing on a few modifications, to jointly use Sigaba for the encryption of messages requiring the highest level of security.

Keyspace

Here is an excerpt from the article "A Practical Meet-in-the-Middle Attack on SIGABA" by George Lasry.

Assuming that there is a set of ten rotors from which the cipher and control rotors are selected, there are 10 ! possible selections for those rotors. Each one of those rotors may be installed in either a forward or reverse orientation. The size of the keyspace for the settings of the ten rotors of the cipher and control banks is therefore:

\( \ \ \ \ 10 ! · 210 · 2610 = 2^{78.8} \)

There are 5 ! possible ordering of the index rotors. The size of the keyspace for the settings of the rotors of the index bank is therefore:

\( \ \ \ \ 5 ! · 105 = 2^{23.5} \)

The combined size of the SI GABA keyspace is:

\( \ \ \ \ 2^{78.8+23.5} = 2^{102.3} \)

However, the size of the keyspace for the index bank is limited by the fact that the five rotors implement a (stationary) permutation of the ten inputs and the ten outputs are mapped by the index output logic into only five outputs. Therefore, the size of the practical keyspace for the index rotors including the index output logic is only:

\( \ \ \ \ 10 ! / 25 = 113,400 = 2^{16.8} \)

and the combined size of the practical keys pace of SI GABA is as follows:

\( \ \ \ \ 2^{78.8+16.8} = 2^{95.6} \)

…

Some characteristics

  • The Sigaba machine has a typewriter type keyboard and prints cipher text or plain text or any text. It can support a typing speed of 45 to 50 words per minute (225 to 250 characters per minute) and a maximum of 60 words per minute (300 characters per minute). It is not recommended to use faster typing because characters may not be taken into account. We see that the machine allows “Offline” encryption. Independently of the machine, it is necessary to transmit the generated ciphertext (by Radio, etc.).
  • The machine measures 15 x 19 x 12 inches (inches) = 38 x 48.3 x 30.5 cm.
  • The machine weighs 93 pounds = 42 kg.
  • Rotors:
    • Rotor 26-contacts: 0.5 in (thickness), 3.5 in (diameter), 5 ounces (weight)
    • Roror 10-contacts: 0.5 in (thickness), 2 in (diameter), 1 ounce (weight)
  • The cost: $ 1,700 (1945) [about $ 30,000 today (2024)]

Production

In total, around 10,000 Sigaba were produced. About 3,300 were used by the Army, the rest by the Navy. Here is the number of machines delivered to the Army: 

     Date           Number  Cumulative
     ---------------------------------
     January 1942     234     234   
     January 1943     373     607    
     February 1944    760   1,367    
     June 1944        788   2,155    
     October 1945    1155   3,310

References

  • The Story of Magic, Memoirs of an American Cryptologic Pioneer, by Frank B. Rowlett, AEGEAN PARK PRESS, 1998. This book recounts Rowlett's career as a cryptanalyst in the US Army and in particular his role in the invention of the concepts which led to Sigaba.

Web Links

  • Wikipedia: SIGABA (link).
  • Electronic Cipher Machine (ECM) Mark II, WHAT IS THE ECM MARK II AND WHY IT WAS IMPORTANT By Rich Pekelney. (link)
  • Crypto Museum (link).
  • Sigaba - ECM Mark II, Jerry Proc's Website. (link).
  • Center for Cryptologic History, The SIGABA / ECM II Cipher Machine : “A Beautiful Idea” by Timothy Mucklow (link)
  • NSA - Friedman documents - History of Converter M-134-C, Volumes 1 & 2, Declassified in 2013. These documents describes the cipher machines created by W.F. Friedman & F. Rowlett (US Army) and whose concepts were precursors to the Sigaba. (Volume 1) (Volume 2)
  • NSA - Friedman's documents - Security of Our High-Grade Cryptographic Systems: link
  • NSA - History of Invention and development of the Mark II ECM - ID: A273704 - link
  • National Cryptologic Museum. Browse the collection (link).
    Note: You can search by keyword. If we use the word "Sigaba", you can recover more than an hundred photographs: the machine exhibited at the museum, a photo of Rowlett showing the Sigaba, the sigaba from every angle, the prototypes of the Sigaba, the rotors, the basket containing the rotors, its different versions , etc.
  • A Practical Meet-in-the-Middle Attack on SIGABA, by George Lasry (link). This article describes among other things, the calculation of the key space.
  • Modern Breaking of Enigma Ciphertexts, by Olaf Ostwald and Frode Weierud (link)