Colossus computer
The Colossus was the first programmable (to a limited extent) digital electronic computer, used for breaking German ciphers in World-War II. It was designed by Max Newman and associates of Bletchley Park, and was built by the British Post Office Research Station at Dollis Hill, by Tommy Flowers and crew.
Colossus was preceded by several computers, many first in some category. Zuse's Z3 was the world's first functional program-controlled computer, and was based on electromechanical relays, as were the (less advanced) Bell Labs machines of the late 1930s (George Stibitz, et al). The Atanasoff Berry Computer of circa 1937 was electronic and arguably the first working digital computer. Assorted analog computers were semiprogrammable, some of these much predated the 1930s (eg, Vannevar Bush). Babbage's Analytical Engine antedated all these (in the mid-1800s), and was both digital and programmable, but was only partially constructed and never functioned. Colossus was the first combining all of digital, programmable, and electronic.
Purpose
It was primarily designed for cryptanalysis in an attempt to break one of the Fish cyphers (a Bletchley Park term) used by the German military for its most secure strategic communications. These were teletype cypher machines in the spirit of that first proposed by Colonel Parker Hitt of the US Army around WWI. The German machines were, essentially, attempts at an electromechanical implementation of the one-time pad cypher invented by Gilbert Vernam (Bell Labs) and Joseph Mauborgne (Signal Corps, USA) in the US at the end of WWI. The most important was a teletype based machine built by Lorenz Electric, the SZ-40 (and later SZ-42) Schlüsselzusatz (meaning, more or less, 'auxiliary key').
Another, different, teletype cypher machine was designed and built by Siemens & Halske, the T-52 Geheimfernschreiber (meaning, 'secret teleprinter'). Early versions of the Siemens machine (the T-52a and T-52b) were used to send signals between Germany and Norway over a cable running through Sweden. The Swedes tapped the cable, copied the traffic, and Arne Beurling, a Swedish mathematician, broke the cypher. Later production versions of the T-52 (there were variants through 'e') were considerably more secure, and quite hard to break even for Bletchley Park. Some of the T-52 traffic was also sent over Luftwaffe Enigma networks which were much more easily broken, and so T-52 traffic was a lower priority for Bletchley Park than might have otherwise been expected.
The one-time pad requires a random sequence. It is combined with the plaintext (bit by bit, usually as character by character) resulting in the cyphertext which is transmitted. On receipt, the same random sequence is combined with the cyphertext (again usually character by character), and because the combining operation is reversible in a particular way (see XOR, for example) the output is the original plaintext. In the German Fish machines, the 'random' sequence was produced by various electromechanical arrangements (on one of them, these were rotors somewhat as in the US SIGABA machine), and the sequence wasn't actually random. Because there were patterns, they could be predicted if the cryptanalysts were sufficiently clever, and plaintexts thereupon recovered. In the case of the Lorenz machine, Col John Tiltman and Bill Tutte of Bletchley Park were sufficiently clever. In the case of the early Siemens machine, Beurling had been sufficiently clever.
Origins
The idea for Colossus developed out of a prior project which produced a special purpose opto-mechanical comparator machine called the Heath Robinson, and its successors the Old Robinson and Super Robinson. The main problem with the Robinsons was synchronising two paper tapes, one punched with the enciphered message, the other representing the patterns produced by the wheels of the Lorenz machine, that tended to stretch when being read at over 1000 characters per second.
Colossus dispensed with the second tape by generating the wheel patterns electronically, and could process 5,000 characters (40 feet / 12m of tape) per second. Colossus Mark 2 was simpler to operate as well as being more advanced, and so greatly speeded the deciphering process, which was largely still carried out by hand. It included the first ever use of shift registers, enabling five simultaneous tests, each involving up to 100 Boolean calculations, on each of the five channels on the punched tape; i.e. up to 12.5 million calculations per second. It was not only able to break the wheel patterns (wheel breaking), but could also determine pin patterns (pin breaking). Both models were programmable using switches and plug panels, in a way the Robinsons had not been.
The project was headed by the mathematician Max Newman. It started early in 1943 and the first version of the machine (Mark 1 Colossus) was finished and installed by about January 1944, to be followed by the improved Mark 2 Colossus in June 1944. Ten Mark 2 Colossus machines were in use at Bletchley Park by the end of the war. Most were destroyed after the war as part of 'protecting secrets' although two survived for many years and were used during the cold war.
Design and operation
Colossus (all versions) used state of the (then) art vacuum tubes (valves), thyratrons and photomultipliers to optically read a cyphertext from a paper tape and applied a programmable logical function to every character, counting how often this function returned "true". Although valves were generally considered to be liable to high failure rates it was recognised that failure occurred at power on and off so the Colossus machines, once turned on, were never powered down until the end of the war.
The Colossus was efficient for its purpose. Even in 2004, Tony Sale notes that "Colossus is so fast and parallel that a modern PC programmed to do the same code-breaking task takes as long as Colossus to achieve a result!".
Whilst Colossus featured limited programmability and was the first of the electronic digital machines to do so, it was not a true general purpose computer, not being Turing-complete. It was not then realized that Turing-completeness was significant; most of the other pioneering modern computing machines were not either (e.g. the ABC machine, the Harvard Mark I electro-mechanical relay machine, the Bell Labs relay machines (by George Stibitz et al), Konrad Zuse's first two designs, and so on). The notion of a computer as a general purpose machine, and not simply a massive calculator devoted to solving difficult but single-minded problems, did not become prominent until many years later.
Influence
The use to which the Colossi were put was of the highest secrecy, and the Colossus itself was highly secret. It therefore had little influence on the development of later computers; being unknown. EDVAC was the early design which had the most influence on subsequent computer architecture. Colossus documentation and hardware were classified from the moment of their creation and remained so after the War.
After the war it is said that Winston Churchill specifically ordered the destruction of the Colossus machines into 'pieces no bigger than a man's hand' and that Tommy Flowers personally burned blueprints in a furnace at Dollis Hill. However two machines continued in use after the war at GCHQ in Cheltenham, until their destruction in the 1960s.
Information about Colossus emerged publicly in the late 1970s after the secrecy imposed by the Official Secrets Act ended in 1976. Thus, Colossus could not be included in the history of computing hardware for many years. Newman and his associates also were deprived of the recognition they were due.
A 500 page technical report on Colossus and Colossus II, entitled General Report on Tunny was released by GCHQ to the national Public Record Office in October 2000; a section is available online .
Reconstruction
In May 2004, the construction of a replica of a Colossus Mk II was completed by a team led by Tony Sale. It currently is on display in the Bletchley Park Museum in Milton Keynes, Buckinghamshire.
See also
References
- Harvey G. Cragon, "From Fish to Colossus: How the German Lorenz Cipher was Broken at Bletchley Park", July 2003. ISBN 0974304506.
- Tony Sale, The Colossus Computer 1943-1996: How It Helped to Break the German Lorenz Cipher in WWII, October 1998. (20 pages). ISBN 0947712364.
