For many years history books named ENIAC as the first large electronic computer. Not out of bad faith, but because another machine, which went into service almost two years earlier, was protected by military secrecy. It was called Colossus, it was built in Britain during the Second World War, and its job was to decipher the communications of the German high command.
Not just Enigma
When people mention Bletchley Park, the British codebreaking centre, they immediately think of Enigma, the cipher machine used by many German units. Colossus, however, had nothing to do with Enigma. It was built for a different and more sophisticated system: the Lorenz cipher machine, used for teleprinter messages between the top levels of the German military. At Bletchley Park the Lorenz traffic was codenamed "Tunny".
How the Lorenz cipher worked
Teleprinters transmitted each character using a five-bit code. The Lorenz machine generated a pseudo-random sequence of characters using twelve toothed wheels and combined it with the plaintext through an operation we would today call XOR. The recipient, with a machine set up the same way, repeated the operation and recovered the original text.
It is the same principle as many modern stream ciphers: security depends entirely on the unpredictability of the generated sequence.
Bill Tutte's achievement
The British had never seen a Lorenz machine. In 1941 a German operator made a serious mistake: he transmitted a long message twice with the same setting, making small changes to the text the second time. Thanks to this error the codebreakers managed to recover a long stretch of the sequence generated by the machine.
The mathematician Bill Tutte analysed that sequence and, working by hand for months, reconstructed the logical structure of the machine: how many wheels it had, with what periods and how they interacted. It is regarded as one of the greatest cryptanalytic achievements of the war.
From method to machine
Knowing the structure was not enough: to read each message, the initial setting of the wheels, which changed, had to be discovered. Tutte devised a statistical method for doing so, but it required an enormous amount of counting. The group led by the mathematician Max Newman built a first machine to automate it, nicknamed "Heath Robinson". It worked, but it was slow and suffered from synchronisation problems between two paper tapes.
Tommy Flowers
At this point Tommy Flowers came on the scene, an engineer of the General Post Office at the Dollis Hill research station in London. Flowers had experience with vacuum tubes, or valves, in telephone exchanges and proposed a radical solution: generate electronically, with valves, the data that Heath Robinson read from a second tape. That left only one tape, the intercepted message, read optically at high speed.
Many doubted that a machine with so many valves could be reliable. Flowers knew that valves failed mostly when switched on and off, and that if left running they lasted a long time.
Colossus in service
The first Colossus, with about 1,600 valves, went into service at Bletchley Park in early 1944. An improved version, the Mark 2, with about 2,400 valves, was operational from June 1944, a few days before the Normandy landings. By the end of the war ten were in operation.
| Feature | Colossus |
|---|---|
| Technology | Vacuum tubes (valves) |
| Input | Paper tape read optically |
| Programming | Switches and plugs |
| Function | Statistical counts on setting hypotheses |
| Output | Electric typewriter |
What it could do
Colossus was not a universal computer: it did not store programs and could not be used for arbitrary calculations. It was, however, programmable, in the sense that its operators, many of them women of the Women's Royal Naval Service, could configure it with switches and plugs to carry out different kinds of logical tests and counts. It is regarded as the first large-scale programmable electronic digital computer.
The long silence
At the end of the war most of the Colossus machines were dismantled and the documentation destroyed or classified. Those who had worked on them could not talk about it, not even with their families. Flowers received no adequate public recognition for decades. Only from the 1970s did information begin to be made public, and further documents were declassified in later years.
Secrecy had a concrete effect on the history of computing. The Colossus experience could not be cited publicly, even though some of the people involved, such as Max Newman, went on to contribute to Britain's first post-war computers.
Between the 1990s and the early 2000s a team led by Tony Sale rebuilt a working Colossus, now on display at the National Museum of Computing at Bletchley Park.
A lesson for security practitioners
The story of Colossus contains a lesson that still holds. The Lorenz machine was well designed, but its security was compromised by an operational error: reusing the same key. In modern cryptography, reusing a nonce or a key in a stream cipher is still one of the most serious mistakes, for exactly the same reason.
In short
- Colossus was built to break the Lorenz cipher of the German high command, not Enigma.
- Bill Tutte reconstructed the structure of the Lorenz machine without ever seeing it, thanks to a German operator's mistake.
- Tommy Flowers designed Colossus using vacuum tubes; the first machine went into service in early 1944.
- Military secrecy hid its existence until the 1970s; a working reconstruction is now at Bletchley Park.
