Rebuilding Colossus

[Sionyn]

great piece on unearthing the secrets to rebuild Colossus now if only dave could sneak in to take it apart
http://www.bbc.co.uk/news/technology-21384672

[G7PSK]

What is still hardly ever mentioned is that the Colossus computer was entirely due to Tommy Flowers, Alan Turing did not believe that  electronic valves would be rugged or reliable enough and wanted the new machines to continue with the use of step relays as the previous one and as used in telephone exchanges. Tommy flowers knew however that if valves were not frequently turned on and off or moved around as in the case of radio's of the day they hardly ever failed and he had been using them with ever greater success in telephone exchanges. He built the first colossus against the wishes and orders of his Superior and Alan Turing.

[Sionyn]

indeed jerrmy clarkson was a big promoter of flowers

[Mechatrommer]

why would anyone want to build an ancient stuffs? do they want to sell it to museum? move on fellas be modern. if you still want to do it, go ahead take a picture, print it big stick it on the rooftop where you sleep, its historic!

[BurtyB]

why would anyone want to build an ancient stuffs? do they want to sell it to museum? move on fellas be modern.

Due to the background of Bletchley Park most of the stuff they have there is "ancient" - it is also the home of The National Museum of Computing and well worth a visit IMHO .

BB.

[baljemmett]

why would anyone want to build an ancient stuffs? do they want to sell it to museum?

Er.  Something like that, yes.  It's ever so slightly historically significant, you see...

(For anybody interested, the National Museum of Computing at Bletchley -- where the reconstructed Colossus resides -- is highly recommended if you're ever nearby!)

[G7PSK]

One reason to rebuild colossus is to see how it works. Churchill was so concerned about secrecy of the code breaking that after the war was over he ordered that all the machines and equipment to be destroyed in case it would be required at a later date, the same sort of political idiocy that made the US declare that atom bomb was secret even to the British who of course had helped build it and therefore knew all about it in any case.

[Neilm]

indeed jerrmy clarkson was a big promoter of flowers

 
wow - he does have some redeeming features..

Also, I believe that Flowers actually paid for and built the original as a proof of concept.

Neil

[smashedProton]

S the colossus is the us designed enigma code breaker?   I think that it uses the guess and check method to determine the plug and rotor positions.  Very interesting.   Does anyone know if it does more than brute force?  I could probably do the same thing with an arduino.   

[SeanB]

True, it did guess by brute force. But at the time it was a very good machine as it was orders of magnitude faster than a human doing the work. It did reduce the amount of work required, as it was able to take a message that was encrypted and by brute force compare it to a known plaintext message fragment that was always present, the headers were consistent across all message types. A common fault with all crypto in that if you know a portion of plaintext you reduce the search space considerably by only looking for keys that produce that output. That, along with the machine being unable to output the same letter or number as encrypted output ( was not possible to do so by the design) reduced the search space by 1/40 and gave a small margin for sanity checking the output, along with getting captured machines to determine the exact operating methods.

70 odd years later and Moore's law has not done much to reduce that, the computers got faster and the crypto got longer, so both ends kept up.

[Neilm]

S the colossus is the us designed enigma code breaker?   I think that it uses the guess and check method to determine the plug and rotor positions.  Very interesting.   Does anyone know if it does more than brute force?  I could probably do the same thing with an arduino.   

I believe that the Colossus was used to decrypt the the Lorenz system not Enigma. Enigma was cracked using the Bombe which was electro-mechanical. The Lorenz system was based on teletype systems.

Neil

[smashedProton]

True, it did guess by brute force. But at the time it was a very good machine as it was orders of magnitude faster than a human doing the work. It did reduce the amount of work required, as it was able to take a message that was encrypted and by brute force compare it to a known plaintext message fragment that was always present, the headers were consistent across all message types. A common fault with all crypto in that if you know a portion of plaintext you reduce the search space considerably by only looking for keys that produce that output. That, along with the machine being unable to output the same letter or number as encrypted output ( was not possible to do so by the design) reduced the search space by 1/40 and gave a small margin for sanity checking the output, along with getting captured machines to determine the exact operating methods.

70 odd years later and Moore's law has not done much to reduce that, the computers got faster and the crypto got longer, so both ends kept up.

Like a German weather announcement

[SeanB]

That, or that all messages began with From and To, and ended with a standard salutation.

[smashedProton]

Hail, mien fuhrer!

[Sionyn]

yes Lorenz it was used by the highest in command  Hitler and his goons

[mark-r]

The major breakthrough decoding the Lorentz machine was when an operator sent nearly the same message twice with the same key. The operator sent the message, the receiver sent a plaintext message asking for it to be repeated, and the operator sent it again almost word-for-word but with small changes. That one mistake allowed the Bletchley Park team to reverse engineer the machine's operation without ever having seen one.

[Mechatrommer]

we got 128 bit SSL 256 bit AES encryption and what not. go crack that for god sake . well, to make it more clear of my previous post, i was not talking particularly on this colossus machine, but also generally on other ancient stuffs that usually popped out in this forum. but no offense guys, its just my 2cnts flavor.

[smashedProton]

Does anybody know the ins and outs of public key cryptography?  Any good resources?  The wikipedia article is pretty unintelligible.  I need some good bedtime reading 

[mark-r]

Does anybody know the ins and outs of public key cryptography?  Any good resources?  The wikipedia article is pretty unintelligible.  I need some good bedtime reading 

This is a pretty good book:

http://simonsingh.net/books/the-code-book/the-book/

[Sionyn]

smashedphoton i done some cryptanalysis for clients, find this book invaluable http://www.schneier.com/book-applied.html
as for  easy explanation for pkey crypto try this http://aolradio.podcast.aol.com/sn/SN-034.mp3

Mechatrommer i dont think we'll ever see a discoverable flaw in Rijndael that what gives AES its strength love beautiful math http://en.wikipedia.org/wiki/Rijndael_key_schedule

if your interested in cryptography Neil Stevenson Cryptonomicon is a very good introduction to Cryptography wrapped up in a awesome story, seriously a very good book that also serves as a eduction in crypto and crypto history.
http://en.wikipedia.org/wiki/Cryptonomicon