14 November, 2005

British rotor machine: Singlet (BID/60)

It's fun to see more information on rotor machines entering the public domain. For example, the Swiss NEMA machine was declassified in 1992. More recently, lots of information on the Soviet rotor machine Fialka (M-125) has been published on the Internet. There's even been a loosening of information about the US/NATO KL-7 system.

This year, a British rotor machine named Singlet (BID/60) was put on display at Bletchley Park in the superb Enigma and Friends exhibit put together by David White and John Alexander. There doesn't seem to have been much -- if anything -- written about this machine in the open literature.

The caption at Bletchley Park reads: "Singlet was used mainly by the British intelligence services C. 1949 / 50 onwards. This is a `Cold War' machine using wired rotors to achieve secure messages. We are very grateful to the Foreign and Commonwealth Office and GCHQ for this opportunity to show `Singlet' here at BP."

Singlet has windows and stepping levers for ten rotors. The rotor tube appears to be a detachable section, labelled BID/60/3, while the base unit is labelled BID/60/1. There is a hint of a connection to the KL-7 in this naming. According to George Mace on Jerry Proc's KL-7 page, the KL-7 components were originally labelled as follows: "the base unit was AFSAM 7/1 (aka KLB), rotor stepping unit AFSAM 7/2 (aka
KLA) and rotor basket AFSAM 7/3 (aka KLK)." The rotor tube, stepping levers and the keyboard are also all somewhat suggestive of some sort of link or common ancestry with the KL-7.

Some photos can be found on Wikipedia/Commons: 1, 2, 3, 4 and 5.

11 November, 2005

"Cryptography is not Sudoku"

In 2002, David Kahn gave a speech to an NSA audience arguing that "cryptanalysis is dead". However, he went on to say that, "though traditional cryptanalysis may be dead, and may have been mostly a corpse for half a century, other opportunities, perhaps more opportunities, lie ahead." Kahn suggested increased opportunities for interception, side-channel attacks, placing back-doors in exported equipment and software, and so forth.

Maybe Kahn is right. I toy around with the following pet conspiracy theory: the recent promotion of codebreaking puzzles and competitions on the NSA and GCHQ websites is misdirection, and an indication that these agencies aren't actually doing a lot of pure cryptanalysis any more. Why would a codebreaking agency call attention to the fact that it breaks codes? Well, a target nation could think, "NSA and GCHQ are about using mathematics to read ciphers. They might still be breaking the systems of [insert developing country here], but our methods are mathematically impregnable, so we don't need to be concerned". Meanwhile, GCHQ or the NSA are compromising their communications through other means.

Anyway, enough with the tin-foil hat stuff. My apologies. Moreover, there's the occasional rumour that cryptanalysis is actually alive and well, as in the recent case of Iraqi leader Ahmed Chalabi. Wikipedia relates the story:

"In June 2004, it was reported that Chalabi gave U.S. state secrets to Iran in April, including the fact that one of the US's most valuable sources of Iranian intelligence was a broken Iranian code used by their spy services. Chalabi allegedly learned of the code through a drunk American involved in the code-breaking operation. Chalabi has denied all of the charges."

Who knows what lies behind this story? Still, a couple of days ago, a reporter speculated that Chalabi broke the codes himself, as he has a doctorate in mathematics. I very much enjoyed the incredulous comments given at the Orbus Quintus blog by "badgerminor" in response.

Sudoku is great fun, by the way.

08 November, 2005


"We're the CryptoKids and we love cryptology. What's cryptology? Cryptology is making and breaking codes. It's so cool."

This is a little weird.

The NSA have a new kids site featuring the "CryptoKids", cartoon animals each with a passion for a different aspect of the fun world of SIGINT and Information Assurance.

The gang includes Crypto Cat (speciality: cryptography), Decipher Dog (cryptanalysis), Joules (engineering), Slate (mathematics), T. Top (computer science), Rosetta Stone (language analyst) and Sergeant Sam (Central Security Service). Each character has their own biography, and there's tons of games and activities: codes and ciphers, things to make at home, colouring pages, online games, recruitment information...

...wait...recruitment information? Oh yes. On the home page, there's a large button that says, "How can I work for NSA?" After all, "it's never too early to start thinking about what you want to be when you grow up". In fact, the entire thing seems to be something of a get-'em-while-they're-young recruitment project: the slogan is "America's CryptoKids: Future Codemakers and Codebreakers".

Don't forget, kids, "without NSA/CSS, American leaders wouldn't be able to talk to one another without the bad guys listening and they wouldn't be able to figure out what the bad guys were planning".

Personally, I would have thought that few kids old enough to be seriously contemplating a future in intelligence work would be young enough to still enjoy the capers of a gang of anthropomorphised cartoon animals (and a gang geekier than your average chess club, to boot), but I'm not complaining...I've got colouring-in to do!

RSA-640 factored

One of the RSA challenge numbers, RSA-640, has been factored:
RSA-640 = 3107418240490043721350750035888567930037346022842727545720161948823
Factor 1: 1634733645809253848443133883865090859841783670033092312181110852389
Factor 2: 1900871281664822113126851573935413975471896789968515493666638539088
The number, which carries a US$ 20,000 prize for its solution, was factored by F. Bahr, M. Boehm, J. Franke and T. Kleinjung using GFNS. The computation took 5 months on eighty 2.2 GHz Opteron CPUs.

RSA-640 is not the largest challenge number to be factored so far -- RSA-200 is larger (despite the confusing name, RSA-200 is 663 bits long, compared to RSA-640, which is 640 bits). RSA-200 was factored in May 2005 by the same team.