Difference between revisions of "Colossus Computer"

From Dead Media Archive
Jump to: navigation, search
(Body of the Colossus)
(How Colossus Worked)
Line 112: Line 112:
  
 
(Sale 357)
 
(Sale 357)
 +
 +
[[File:Block Colossus Diagram.jpg|200px|thumb|right|Block Diagram of Colossus]]
  
 
In brief, the Colossus read intercepted, encoded text at 5,000 cps. The Colossus would 'read' the text multiple times and use a complicated Boolean function between the text and the wheel patterns used to encode it. Boolean functions are the basis of computers, registering results as either true or false based on the punched holes in the tape.
 
In brief, the Colossus read intercepted, encoded text at 5,000 cps. The Colossus would 'read' the text multiple times and use a complicated Boolean function between the text and the wheel patterns used to encode it. Boolean functions are the basis of computers, registering results as either true or false based on the punched holes in the tape.
  
 
The Colossus was made up of, basically, an optical reader system, a master control panel, thyratron rings and driver circuits, calculators, shift registers, logic gates, counters, and printer logic (Sale 357).
 
The Colossus was made up of, basically, an optical reader system, a master control panel, thyratron rings and driver circuits, calculators, shift registers, logic gates, counters, and printer logic (Sale 357).
PICTURE: SALE 359
 
  
 
== Shadow of the Colossus ==
 
== Shadow of the Colossus ==

Revision as of 15:44, 11 November 2010

The Colossus is arguably the first digital computer in human history. Created during World War II in order to help decipher encrypted German messages sent by the Lorenz Cipher, it stands as both an influential, technological advancement as well as a singular creation unto itself. Although its creation spawned a technological revolution in terms of electronic computation, its existence remained secret for over 30 years. Unlike the majority of technological media, the Colossus Computer was built with a singular purpose in mind, achieved its goal, and was immediately destroyed. It was never re-appropriated for a new use, except perhaps as an historical object, as it has recently been 'restored' as part of a museum for World War II code breaking at Bletchley Park.

In this way, as a technological medium, the Colossus is both hugely influential and entirely unique. Its role as a translator of another medium, that of German code, is a position not seen easily today. Perhaps this is why the Colossus can be easily forgotten, but deserves acknowledgement and remembrance.

Colossus Computer with two WREN operators

Life of the Colossus

Machine-Cryptography Before Colossus

Following WWI, cryptography was becoming more mechanized than ever before. Along with this trend came a development of cryptanalysis as its own field of science and study. Since these new enciphering schemes were becoming more and more mechanical. the process of breaking their codes was also turning into a more electronic field. This development would not only change the course of cryptology, but begin the field of computer science that continues to this day. (Ratcliff 198)

The Bombe

The first German cipher to utilize mechanical means was the Enigma machine. Code-breaking took its first steps into the mechanical realm with the development of the Bombe, developed by a primarily Polish team of cryptographers and engineers. The concept was simple: electricity utilized by a machine would allow for faster checking of possible decipherments than human checking. It should be noted that using machines to aid in deciphering was used before the Enigma. Germans and Americans relied on a punch card-based system to sort information. The Bombe, however, worked electromagnetically, like the Enigma it was attempting to break. This meant it, in many respects, replicated the machine itself. Rather than just checking possibilities for the message to be deciphered, the Bombe aimed to use the design of the encoding device itself, the Enigma machine, to decode the messages. Developed in the late 1930s, the Bombe and its reaction to the German Enigma machine would serve as a major influence when the time came for a new cipher to be broken, requiring an even more complicated and sophisticated machine to crack it. (Ratcliff 203-205)

A New Cipher, A New Problem

The aim of the cryptographer is to produce a system that is convenient or simple to use legitimately but is too expensive, complex, or impossible in principle for the cryptanalyst to break ~ Jack Good (Good 149)

Lorenz SZ42

After the German Enigma machine was effectively proven useless due to Allies breaking the code, a new encoding devised was devised known as the Lorenz SZ 40/42, named after the Lorenz Company that was commissioned to create the device. This device would allow for secure radio communication based on the 'additive method' of enciphering invented in 1918 by Gilbert Vernam. This meant the Lorenz utilized a non-Morse cipher known as Baudet (Ratcliff 205).

The Vernam method was simple enough; a text would become enciphered by adding obscure and 'randomly' generated characters to the actual textual message. The receiver of the message would then use their own Lorenz machine to add the same obscuring characters (by setting their rotors to the correct starting position for said-message), and cancel them out, revealing the intended message. This entire method of addition was known as Modulo-2 addition, referencing the addition of characters on both ends of the communications line (Sale 351).

The key to both the perceived success and ultimate failure of the Lorenz cipher system was the concept of creating random, obscuring characters. If the characters were created randomly by a computerized machine, the code would be unbreakable without a corresponding machine set to the same positions as the sender. Theoretically, this meant the Lorenz cipher system was a perfect example of encoded communication, as only those with the proper technology would be able to gain access to the intended message/text.

Lorenz SZ40/42

In brief, the cryptographic machine converts input (plain language, P) into a cipher (Z) by using some function (f). The equation is thus, Z=f(P,K), with K being the key to the cipher. Without K, presumably, Z cannot be converted into P by means of the function. If K has N possible values, the higher N is, the more difficult the code is to break. With a random character generator, N has a very high value. (Good 150).

However, the machination of a random character generator also reveals the system's brutal flaw; there is no such thing as a random character generator. What is actual made is a "pseudo random sequence" of characters. This would eventual lead to the Bletchley Park code-breakers to be able to truly break the Lorenz cipher without requiring access to a Lorenz machine itself (which would not have been breaking the code as much as stealing it). In fact, the code-breakers at Bletchley Park "never saw an actual Lorenz machine until right at the end of the war, but they had been breaking the Lorenz cipher for two and a half years" (Sale 352).

Fish and Tunny: Breaking the Code

It was 1940 when the first Lorenz-based German transmissions were intercepted by Allied officials in the UK. This non-Morse traffic was given the code name "Fish," a reference to the fact that this system of transmission was known by some Germans as Sagefish (Hinsley 141). "Fish" was clearly generated by a new machine, as the Enigma was based in Morse. What was also clear was the fact that the Germans based the transmission system around the concept that the same machine was used to both encipher and decipher the message over the radio (Hinsley 141).

"Tunny" was the word used to describe the "Fish" traffic messages as well as the machine that was being used to encipher/decipher it. The Tunny machine had two sets of five coding wheels or rotors that were used to generate the obscuring characters. In order to break the code and convert the cipher into plain language text, the receiver had to know the starting positions of the code wheels (Halton 170).

In code-breaking, a 'Depth' is more than one message, making it possible to hopefully see patterns and break the code. As more and more Depths were accrued through 1941, the hope of breaking the code did not seem to brighten. However, a critical German error allowed for British cryptanalysts to make the biggest break yet. On August 30, 1941, a nearly 4,000 character piece of text was sent between two German Lorenz machines and intercepted by British agents. The receiving end of the message sent back a transmission by radio a request for the message to be sent again, as (for whatever reason) it had not been properly received. The sender and the receiver both reset their Lorenz machines to the same starting position as the last transmission. Now, the German sending the message typed this second text slightly different from the first, this time making abbreviations and short-hamnd at points, clearing not wanting to type the entirety of the previous message again. The resulting text was about 500 characters shorter than the first. Thus, two mistakes were made. First, the Germans reset their wheels to the original position, meaning both texts would use the same obscuring numbers (pseudo randomly generated). Second, the second message was typed out differently. If it was the same, the same ciphered text would have been intercepted, and would have not been able to be broken. However, since the texts were different, but using the same obscuring characters, a pattern would become more visible. (Sale 353)

A young recent graduate at Bletchley Park named Bill Tutte was given the texts and, without a computer, started working out how the code must have been generated. Over the next two months, the structure of the Tunny machine was worked out. It cannot be oversold how amazing this was as a code-breaking effort. How they were able to deduce not only the ciphers but how the machine created them is astounding. However, a major issue was still noting where the code wheels had to be in the starting position. In essence, this was simply a "guess-and-see" game that could take anywhere between four to six weeks to discover. This was too long, as the messages deciphered by then were worthless. It was at this point that the machine was meant to step in as an answer. (Sale 354)

Rise of the Machines

Because the code took too long to break by hand, it was determined that a speedy, mechanized device was the only way to help break the code in a reasonable amount of time. The first electronic machine built to help break the "Tully" was the Heath Robinson, named after the British cartoonist of fantastical machines (a British equivalent of Rube Goldberg). The biggest problem with the Heath Robinson was its difficulty in synchronizing two pieces of input tape to read at 1,000 characters per second (cps). One tape had the Lorenz wheel patterns on it, the other the encrypted text. Both had to run equally multiple times through the machine in order to determine the starting position of the Lorenz wheels. While Heath Robinson worked in theory, enough to show Max Newman's mathematical theory of applying code-breaking to this electronic machine was accurate, it was too problematic to be effective. (Sale 354)

PICTURE: COPELAND 31

The Colossus fixed this problem by having the wheel patterns be generated electronically within the machine itself, fixing the need for synchronization. This massive intellectual contribution by Tommy Flowers led to the creation of the Colossus Mk 1. The Colossus was up and running by the end of 1943. The result was an increase in the speed of decoding messages from weeks to hours. The Colossus aided in several Allied programs, most notably in the lead up to D-Day. Intercepted messages deciphered by Colossus proved that Hitler and the Germans were falling for the deception campaigns laid out by the Allies leading up to the Normandy invasion. This gave the Allies confidence in launching the crucial assault and bringing an end to the war. (Sale 355)

Colossus Wiped From Memory

Over the course of the war, Colossus was upgraded to the Colossus Mk 2 in June of 1944. By the end of the war, there were ten functioning Colossi in the Bletchley Park facilities. Following the war's end on Victory Day, the Colossi were deemed to have no more function and were ordered to be dismantled immediately. Eight were dismantled, while the other two were sent back to the Government Code Headquarters until being destroyed in 1960. All the drawings and schematics were burned and the entire existence of Colossus was kept secret. It was not until the 1970s that information began to emerge. In the 1980s, some of the researchers and scientists behind its invention began writing papers on the Colossus, its function, and its contributions to not only the war effort, but to the invention of the computer. (Sale 362)

Mind of the Colossus

The Mansion at Bletchley Park, Headquarters of British Code-breaking

The men and women who created and operated the Colossus are many, yet specific. Unlike many technologies with varied origins, the creation of the Colossus computer was done with keen intent and diligence by specific, known people. They came from many walks of life and were motivated by the hopes of serving their country in a wartime effort. The early code-breaking computers, thus, "sprang from the imaginative, sometimes desperate, often improvised innovations of the Anglo-American war effort against German machine ciphers" (Ratcliff 198). People came from many fields, both theoretical and practical, in order to help. In this way, the construction of the Colossus was motivated by nationalism more than any believe in scientific progress. The following are just a few of the key figures in the creation of the Colossus and the breaking of the Tully.

Max Newman

Max Newman

Max Newman was in charge of the Tunny-breaking team at Bletchley Park. This group would come to be named after him, the 'Newmanry.' The 'Newmanry's' function was "to work on machine attacks on Tunny, and it complemented the Testery, where hand and linguistic attacks were used" (Good 160). He was the main managerial force behind the creation and overseeing of Colossus (and the Heath Robinson before that). He deserved credit in the creation of Colossus mainly in this administrative capacity, as well as deciding to bring in the brilliant Tommy Flowers who, it will be discussed, was the main creative force behind building Colossus (Good 163).

Jack Good

Jack Good

Jack Good was a statistician summed to Newman's team to help build an electronic machine to aid in the decoding process. He had previously worked on the Enigma code-breaking and the construction of the Bombe, as well as the Colossus's immediate predecessor, the Heath Robinson machine.

According to Good, "one of the greatest secret inventions of the war was the discovery that ordinary teletype tape could be run at thirty miles per hour without tearing" (Budiansky 314). In many way, Good helped pave the way for mechanized aid in code-breaking. It wasn't until he received help from TommyFlowers that the Colossus would take form.

Bill Tutte

Bill Tutte

Bill Tutte is the man who broke the Tunny code before any machines were even involved. Using two streams of intercepted text, Tutte was able to figure out the entire structure and functionality of the Tunny machine (Lorenz Cipher) without ever having seen it. Granted this took months, and individual decipherings could possibly take weeks, making the translated text unhelpful. That is why the creation of the Colossus was taken up (Good 161).

Tommy Flowers

Post Office Research Station at Dollis Hill

Tommy Flowers was a British Post Office electronics engineer working at the Research Station at Dollis Hill. He was recruited as an automaton specialist by the Government Code and Cipher School (GC&CS). Along with Jack Good, he worked on theories derived by noted mathematicians Alan Turing and J von Neumann that led to the production of the Colossus (Ratcliff 205).

The biggest contribution to the Colossus construction from Tommy Flowers was the suggestion that the wheel patterns in the machine be generated electronically. In this way, there was no need for dual paper tape, and the issue that plagued the Heath Robinson machine, synchronization, was eliminated (Sale, 354). "Alan Turing contributed to the thinking in developing these machines, as did Max newman and several others, but an enormous part of the credit for designing Colossus, and all the credit for building it, goes to Tommy Flowers" (Copeland 192).

Body of the Colossus

The Eight Racks of the Colossus

Structure

In total, there ended up being ten Colossi computers, the original Mk 1 and Mk 2, and eight replications of the Mk 2. The Colossus was made up of eight racks, arranged in two rows of four. The racks were 90 inches high. Each bay, holding four racks, stretched about 16 feet across. In addition to the racks, there would be a paper tape handler at one end as well as a output typewriter at the other (Sale 355).

IBM Electromatic Typewriter

The typewriter was a modified version of an IBM Electromatic Typewriter, common in the 1940s. These modifications, placed by Tommy Flowers and his team, allowed the typewriter to interface with one of the Colossi. The typewriters were capable of printing ten characters per second, a high rate of performance that was necessary to quickly complete deciphering (Cragon 99).

How Colossus Worked

Input: Cipher text punched onto 5-hole paper tape, read at 5,000 characters per second (cps)

Output: Buffered onto relays, typewriter

Processor: Memory of 5 characters of 5-bits held in a shift register, pluggable logic gates, 20 decade counters arranged as 5 by 4 decades

Clock Speed: 5 KHz, derived from sprocket holes in the input tape

Vavles: 2500

(Sale 357)

Block Diagram of Colossus

In brief, the Colossus read intercepted, encoded text at 5,000 cps. The Colossus would 'read' the text multiple times and use a complicated Boolean function between the text and the wheel patterns used to encode it. Boolean functions are the basis of computers, registering results as either true or false based on the punched holes in the tape.

The Colossus was made up of, basically, an optical reader system, a master control panel, thyratron rings and driver circuits, calculators, shift registers, logic gates, counters, and printer logic (Sale 357).

Shadow of the Colossus

Rebuilding

The Colossus rebuild project was begun by Anthony Sale and colleagues in July, 1994 as part of the opening of the Museums of Bletchley Park dedicated to the history of code-breaking in World War II. Some actual Post Office and radio engineers ended up taking part in the project. A basic functioning Colossus was finished by June of 1996, with Dr. Tommy Flowers in attendance for the 'switch-on' occasion (Sale 362).

The major purpose of the rebuilding of the Colossus and the opening of the Museums at Bletchley Park is an attempt at changing the conception that the American-made ENIAC was the first digital computer ever made. In fact, the Colossus was completed two whole years before the ENIAC was switched on in 1944, yet the knowledge and nature of its secret war origins and the immediate destruction of the devices following their use has led to the Colossus being forgotten in computer history. The tide is slowly turning, however, as more information and research is being revealed into the creation and implementation of the Colossus.

Meaning of the Colossus

Bibliography

Andresen, S.L. “Donald Michie: Secrets of Colossus Revealed.” IEEE Intelligent Systems. Vol. 16. No. 6

Budiansky, Stephen. Battle of Wits: The Complete Story of Codebreaking in World War II. New York, NY: The Free Press, 2000.

Cragon, Harvey G. From Fish To Colossus: How the German Lorenz Cipher was broken at Bletchley Park. 2003.

Copeland, B.J. “Colossus: Its Origins and Originators.” Annals of the History of Computing, IEEE. Vol. 26. No 4.

Copeland, B.J. Colossus: The Secrets of Bletchley Park’s Codebreaking Computers. Oxford University Press, 2006.

Good, Jack. “Enigma and Fish.” Codebreakers: The Inside Story of Bletchley Park. Eds. F.H. Hinsley & Alan Stripp. Oxford Univeristy Press, 1993.

Halton, Ken. “The Tunny Machine.” Codebreakers: The Inside Story of Bletchley Park. Eds. F.H. Hinsley & Alan Stripp. Oxford Univeristy Press, 1993

Hayward, Gil. “Operation Tunny.” Codebreakers: The Inside Story of Bletchley Park. Eds. F.H. Hinsley & Alan Stripp. Oxford Univeristy Press, 1993

Hinsley, F.H. “An Introduction to Fish.” Codebreakers: The Inside Story of Bletchley Park. Eds. F.H. Hinsley & Alan Stripp. Oxford Univeristy Press, 1993.

Ratcliff, R.A. Delusions of Intelligence: Enigma, Ultra, and the End of Secure Ciphers. New York, NY: Cambridge University Press, 2006.

Sale, Anthony. “The Colossus of Bletchley Park.” IEEE Review. Vol 41 Issue 2. 1995

Sale, Anthony E. "The Colossus of Bletchley Park – The German Cipher System." The First Computers: History and Architectures. Eds. Raul Rojas and Ulf Hashagen. Cambridge, MA: The MIT Press, 2000.

Wells, Benjamin. “Advances in I/O, Speedup, and Universality on Colossus, an Unconventional Computer.” Unconventional Computation. 8th International Conference, 2009.