█ LARRY GILMAN
A cipher machine is a mechanical device that assists in the production of ciphertext from plaintext and vice versa. In this broad sense, any mechanical aid from a cipher wheel to a supercomputer can qualify as a cipher machine; however, the term is usually reserved for devices that are fairly complex and that operate on mechanical or electromechanical rather than on electronic principles.
Before World War I, ciphers were implemented using either marks on paper or simple aids such as cipher wheels. After the war, a number of inventors in various countries produced cipher machines that transferred the complexity and tedium of ciphering to a mechanism. These machines allowed the operator, who might be completely ignorant of the cipher's nature, to simply type at a keyboard or enter characters one by one by moving a wheel with their fingers. If plaintext (ordinary written language) was entered into such a machine, ciphertext (apparently random characters) was produced; if ciphertext was entered, plaintext was produced. Cipher machines made it possible to cipher and decipher large numbers of messages with less training for personnel, fewer errors, and higher speed.
Many cipher machines invented in the post–World War I period employed as their key component the scrambler disk or rotor. The typical rotor is a disk a few inches in diameter, with letters and numbers printed around its rim and embedded wires connecting one side to the other. Matching points on opposite surfaces of the disk correspond to the same alphanumeric characters, and each wire running through the disk corresponds to one character to be enciphered or deciphered. By connecting one point on surface A of the rotor—say, the point corresponding to the letter M—to a different point on surface B—say, the point corresponding to the letter Z—the rotor implements a fixed substitution cipher (i.e., replaces every character by some other). In this example, M is enciphered to Z and Z is deciphered to M (or vice versa).
The substitution cipher built into the wires of a single rotor is a trivial one. What the inventors of the rotor-based cipher machines realized was that by lining up multiple cipher disks and continually rotating them as a message was enciphered or deciphered, they could produce much more formidable ciphers. For instance, three rotors could be stacked or aligned so that surface B of rotor 1 met surface A of rotor 2, while surface B of rotor 2 met surface A of rotor 3. Each letter of the input (at surface A of rotor 1) then follows a tortuous path through the wiring of all three disks to the output (at surface B of rotor 3). If the rotors are shifted upon encryption or decryption of each and every message character, the encryption/decryption path is not only tortuous, but also changing. A degree of cipher security that was essentially impossible with pencil-and-paper ciphering was made possible by such machines.
The rotor principle was discovered independently by inventors in several countries, the most famous being German engineer Arthur Scherbius (1878–1929). Scherbius invented a three-rotor cipher machine, the Enigma, in 1918 (the last year of World War I). Scherbius tried unsuccessfully to sell his machine to commercial buyers, but he was ahead of his time; corporations did not begin to use encryption widely until the 1960s. Enigma was, however, purchased by the German government in 1926. At that time, Germany was busy rebuilding its military forces after its defeat in World War I and the humiliating terms of the Treaty of Versailles. Furthermore, the German military leadership had become aware that their pencil-and-paper field cipher, the famous ADFGVX cipher, had been broken by French cryptographers only a few months after its deployment in 1918, leading to at least one significant military defeat for the Germans. In order to prevent a repetition of the ADFGVX debacle, the Germans switched to Enigma as their primary system for secret communications.
The different branches of the German military also employed slightly different models of the Enigma cipher machine. In 1943, the German military deployed the SZ42 cipher machine for use over 26 crucial communications links. The SZ42 employed the stream-cipher technique, in which identical key-streams of pseudorandom characters are generated at both the sending and receiving end of the link and added, character by character, to the individual characters of the plaintext (for ciphering) or ciphertext (for deciphering). The German military did not replace Enigma with the SZ42 for general use because the SZ42's complexity made it too heavy for the field.
The SZ42 cipher proved difficult for allied cryptographers to crack, as did another German cipher machine, the Geheimschreiber, first deployed by the German navy in 1942. However, Allied cryptographers cracked the Enigma, SZ42, and Geheimschreiber ciphers by building specialized devices to systematically try out possible keys
for the decryption of messages. The first such devices— "bombes,"invented by Polish mathematician Marian Rejewski (1905–1980) and possibly named for the loud ticking noises they emitted while functioning—were electromechanical (i.e., used a combination of electrical currents and moving parts). Bombes sufficed for the Enigma cipher, but to crack the SZ42 and Geheimschreiber ciphers, the Allies built what is sometimes considered the world's first electronic computer, the Colossus. The Colossus was based primarily on the ideas of British engineer T. H. Flowers (1905–1998) and British mathematician Alan Turing (1912–1954). (An "electronic" computer, as opposed to an electromechanical device, does not use moving parts to perform its calculations.)
Cipher-machine technology reached its peak in the Geheimschreiber and SZ42 cryptosystems, achieving a level of cryptographic security that could only be breached by the invention of a wholly new technology: the electronic computer. Nevertheless, all the major German ciphers of the World War II—and the primary Japanese cipher too, codenamed Purple—were broken by the Allies.
The Allies also used cipher machines during World War II, but with better luck, as the Axis governments did not succeed in breaking Allied ciphers routinely. The United States Army's primary cipher machine descended from a compact device invented by Swedish inventor Boris Hagelin (1892–1983) in the mid 1920s. Hagelin's cipher machine, originally designated the B-21, sold thousands of copies to the French military between 1934 and the French defeat in World War II. The U.S. Army purchased Hagelin's machine after the German invasion of Norway in 1940 and redesignated it the M-209. More than 140,000 M-209s were manufactured before the end of the war. The M-209, like the SZ42, employed the stream-cipher technique, with matched generation of the key-stream at the transmitting and receiving ends of each link. Interestingly, this technique is still used today in applications such as digital pay-TV, file encryption, and communication with secure Web sites; however, electronic, rather than mechanical, generation of the pseudorandom key stream is used.
Cipher machines continued to be used by many countries for some years after the end of World War II, but were slowly rendered obsolete by the increasing availability of general-purpose digital computers. The displacement of cipher machines by computers was inevitable for several reasons. A computer can be flexibly reprogrammed to implement any number of ciphering schemes, whereas a cipher machine can implement only the cipher it is built for. Further, electronic computers operate at far higher speeds than can mechanical devices. Today, all serious ciphering is performed using digital computers, and the only remaining ciphering machines are in museums.
█ FURTHER READING:
Churchouse, Robert. Codes and Ciphers. Cambridge University Press, 2002.
Deavours, Cipher, et al. Cryptology: Machines, History and Methods. Norwood, MA: Artech House, 1989.
Singh, Simon. The Code Book. New York: Doubleday, 1999.