Longevity of Colossus

2 August 1999

From: Paul Leyland <pleyland@microsoft.com>
To: "'ukcrypto@maillist.ox.ac.uk'" <ukcrypto@maillist.ox.ac.uk>
Subject: Historical question: Longevity of Colossus.
Date: Mon, 2 Aug 1999 06:57:51 -0700

I visited Bletchley Park on Friday as part of the Microsoft Research works outing. Tony Sale showed us around and mentioned something that I'd heard before but never thought about. For those who don't yet know, Tony was the chap who rebuilt a Colossus at BP.

Tony said that all but two of the wartime Colossi were demolished, the remaining pair being carted off to Cheltenham where they remained until they were destroyed in 1960 or so.

My question: why did GCHQ keep two of them for so long? If they only had one, it could have been an exhibit in a museum. Keeping two bits of kit, each occupying several cubic meters and requiring 5.5kW of power to run suggests that they were still useful.

Does the answer to this question have any relation to the observation that the design of Colossus was classified until very recently?

Paul

From: Michael Bacon <MBacon@snci.co.uk>
To: ukcrypto@maillist.ox.ac.uk
Subject: RE: Historical question: Longevity of Colossus.
Date: Mon, 2 Aug 1999 16:10:17 +0100

So far as I am aware, decryption of German signals (especially those using the Naval Emigma machine) continued for quite some time after the war. Additionally (heresay), captured Enigma machines were refurbished and sold to other governments.

Streaky

_____
~(_____)>
" "

The opinions given are my own and are not necessarily representative of those held by my employer.

From: Paul Leyland <pleyland@microsoft.com>
To: "'ukcrypto@maillist.ox.ac.uk'" <ukcrypto@maillist.ox.ac.uk>
Subject: RE: Historical question: Longevity of Colossus.
Date: Mon, 2 Aug 1999 08:37:19 -0700

> So far as I am aware, decryption of German signals (especially
> those using the Naval Emigma machine) continued for quite some
> time after the war. Additionally (heresay), captured Enigma
> machines were refurbished and sold to other governments.

While the above is undoubtedly true, and probably explains in large measure why knowledge that Enigma traffic had been comprehensively broken was classified for almost thirty years, it doesn't really answer my question.

Colossus was not used to break Enigma. It was used on the output of the Lorenz machines. These enciphered teletype traffic by exoring with a keystream generated by a rotor machine far more complicated than even Naval Enigma.

Paul

Date: Mon, 2 Aug 1999 17:48:21 +0100 (GMT)
From: Quentin Campbell <Q.G.Campbell@newcastle.ac.uk>
To: "'ukcrypto@maillist.ox.ac.uk'" <ukcrypto@maillist.ox.ac.uk>
Subject: Re: Historical question: Longevity of Colossus.

On Mon, 2 Aug 1999, Paul Leyland wrote:

> My question: why did GCHQ keep *two* of them for so long? If they only had
> one, it could have been an exhibit in a museum. Keeping two bits of kit,
> each occupying several cubic meters and requiring 5.5kW of power to run
> suggests that they were still useful.
>
> Does the answer to this question have any relation to the observation that
> the design of Colossus was classified until very recently?

Brian Randell is probably among the best placed on this list to help answer that.

A couple of things I recall from talks given here:

1. At the end of the War, Donald Michie and others tried "programming" a Colossus to do some non-cryptographic computations. I also understand it was tried on some cryptographic tasks for which it was not originally designed. It may be it had some utility in these other areas that extended its useful life.

2. Tony Sale has also said that there are some analysis algorithms that are built into the logic of Colossus and which could be selected as part of its set-up menu that remain classified still. He was told that some of these techniques were still in use at Cheltenham until fairly recently.

Quentin

PHONE: +44 191 222 8209 Computing Service, University of Newcastle
FAX: +44 191 222 8765 Newcastle upon Tyne, United Kingdom, NE1 7RU.
-------------------------------------------------------------------------
"Any opinions expressed above are mine. The University can get its own."

From: Quentin Campbell

> A couple of things I recall from talks given here:
>
> 1. At the end of the War, Donald Michie and others tried "programming"
>    a Colossus to do some non-cryptographic computations. I also understand
>    it was tried on some cryptographic tasks for which it was not
>    originally designed. It may be it had some utility in these other areas
>    that extended its useful life.

Thank you, that is information I had not heard before.

> 2. Tony Sale has also said that there are some analysis algorithms that
>    are built into the logic of Colossus and which could be selected as
>    part of its set-up menu that remain classified still. He was told
>    that some of these techniques were still in use at Cheltenham until
>    fairly recently.

This ties in nicely with a tale I can relate.

Bob Morris, the retired chief scientist at NSA, gave a talk at Cambridge about 3 years ago. In the pub afterwards he claimed that none of the WW2 machine cryptography was anything but of historical interest. I asked him why aspects of the Colossus design were still classified. He immediately backtracked, giving essentially the same answer as Tony Sale gave to you.

About a year ago, New Scientist (I think) revealed one of those aspects that had recently been declassified. Incorporating a one-bit delay in the keystream (using a single capacitor in Colossus) and then correlating gave a useful attack on a class of stream ciphers and that this trick was not previously widely known.

Paul