5 ms·

The French semaphore system hack of 200 years ago

38 points by Clepsydra 8y ago

crooked-v 8y ago
One of the Discworld novels uses 'hacking' of multi-shutter semaphore towers as a plot point, by way of a gibberish sequence that can physically destroy the poorly engineered towers by slamming the shutters around at a resonant frequency (book) or cause the controls to physically seize up in a jammed position (film).
- smallnamespace 8y ago
  Real physical layers are quite vulnerable to this type of disruption and we actually mitigate this by paying rather high costs at the bit encoding level: https://en.wikipedia.org/wiki/8b/10b_encoding https://en.wikipedia.org/wiki/8b/10b_encoding https://en.wikipedia.org/wiki/64b/66b_encoding https://en.wikipedia.org/wiki/64b/66b_encoding https://en.wikipedia.org/wiki/Scrambler https://en.wikipedia.org/wiki/Scrambler
- sophacles 8y ago
  GNU Terry Pratchett http://www.gnuterrypratchett.com/ http://www.gnuterrypratchett.com/
paxy 8y ago
I'm pretty sure you can go back a lot further if you keep redefining the meaning of the word. Would some Roman general sending a spy into an opposing camp to steal codes be a cyber attack?
- tzar 8y ago
  The semaphore ploy strikes me as much more rooted in information, thus the description “cyber”. Stealing code is just theft, since the object of theft doesn't generally qualify the act (though it does in some cases).
- lainga 8y ago
  It is cyber in the sense of cybernetics ("the mathematical study of communication and control in the animal and the machine")
  - kazinator 8y ago
    Let's see: https://www.etymonline.com/word/cybernetic#etymonline_v_36576 https://www.etymonline.com/word/cybernetic#etymonline_v_3657... "kybernetikos": Greek for "good at steering". Thus military attack involving an excellent steering maneuver is cyber attack.
    - Udik 8y ago
      Not too sure about the -tikos thing and the "good at" part, google (and the Oxford dictionary) gives etymology from kubernetes, pilot. Cybernetics is then, I guess, the science of (automatic) control, so nothing to do with "steering".
- kazinator 8y ago
  Does the Trojan Horse ring a bell?
zitterbewegung 8y ago
A better title would be the first hack of a financial market.
- dang 8y ago
  We've given the article a narrower title above, based on its actual content.
kazinator 8y ago
An inside job to borrow some network resources does not qualify as an attack on a network. This is like using your neighbor's Wi-Fi without permission being a "cyber attack".
- zitterbewegung 8y ago
  It is a government owned network that only allowed government traffic. It’s more like borrowing Siprnet...
jgeralnik 8y ago
My favorite story that I've talked about as the first "cyber" attack for years is from the time of the second Jewish temple, a few hundred years B.C. The Jewish calendar is based on the lunar cycle, and each month has either 29 or 30 days. Back then the length of months was not declared in advance. Instead, each month when a new moon was spotted witnesses could come to the religious court (called Sanhedrin) in Jerusalem claiming to have witnessed the new moon, and if the court accepted their testimony that day was declared the first of the month. To spread the word they would light a bonfire on a mountain, and every city that saw the bonfire would light their own and the message would quickly spread. (Similar to the semaphore system but only transmitting a single bit. Because the signal was just a fire at nighttime it would reach very far very quickly) Now, there was a small offshoot of Jews known as Karaits, whose interpretation of the Bible caused them to believe that passover must fall on a Sunday (the passage says something like "on the first day [of passover] " which can also be interpeted as "on Sunday"). "Mainstream" (at the time) Judaism did not agree with this interpretation. So what did the Karaits do? On the day that they wanted to be the first of the month (such that Passover, the fourteenth of the month, would fall on Sunday) they went to a nearby mountain and lit a bonfire. The other cities saw the bonfire and assumed it was legitimate, and so lit their own bonfires as well, and the fake message was spread. Classic man in the middle attack! So how did the Sanhedrin solve this issue? One way would be to sign their messages with RSA. Generate two large prime numbers p and q, have N=p*q and e=65537 be public, known parameters, figure out how to properly pad your messages so that you're not using a completely broken system, and sign the hash of the message with phi=(p-1)(q-1). There are a few problems with this. First, it's difficult to transmit 1024 bit messages over fire signals. Second - Rivest, Shamir and Adelman wouldn't be born for another 2000 years or so. Instead, the solution was to move to a different, more secure medium. Instead of lighting a fire, the Sanhedrin would send messengers to spread the word to nearby cities, and each city would send their own messengers to their neighbors. Each city would trust their immediate neighbors and a chain of trust could be established. The problem with the new solution is that it is much much slower than the traditional fire signals, so some far off places wouldn't know when the previous month started by the time the 14th of the month (passover) came around. To handle this, far off cities would celebrate two days of passover - one on the 14th assuming the previous month had 29 days, and one on the following day under the assumption that there were 30 days. To this day, the tradition of Jews in far off places celebrating 2 days of holidays remains. Orthodox religious Jews outside of Israel celebrate 2 days of each of the 3 major Jewish holidays out of a sense of "doubt" over which day the holiday actually falls on because the message would not arrive in time.
twiceaday 8y ago
Tom Scott did a video on this: https://www.youtube.com/watch?v=cPeVsniB7b0 https://www.youtube.com/watch?v=cPeVsniB7b0
imglorp 8y ago
> Looking back at the incident now, one can see that the French semaphore system was hopelessly insecure,... Actually, this would have been an ideal application of one-time-pads for end-to-end encryption of the sensitive information, needing no modern tech beyond pencil and paper. The pads could have been passed by hand between participants and would have been highly secure.
- Udik 8y ago
  It seems that messages were actually "encrypted" and "compressed" with a shared code book. But encryption wouldn't have changed anything in this particular case: nobody was altering or decoding the authentic messages, they were just piggybacking on the error correction codes to insert more information. Very clever.
Udik 8y ago
Very interesting. A less clever but very similar attack, certainly inspired by this, is the described in the great classic "the Count of Monte Cristo", written ten years later. The Count manages to ruin one of his enemies by bribing a semaphore operator in transmitting news of a political unrest in Spain, causing a collapse in the value of foreign bonds (if I remember the details correctly).
credit_guy 8y ago
Most likely this event inspired Dumas, who put a similar hack in his Count of Monte Cristo novel. It’s funny that Dumas’s hack is less sophisticated that the one done in real life, but the whole chapter is quite well written. It’s available online at [1]. [1] https://en.m.wikisource.org/wiki/The_Count_of_Monte_Cristo/Chapter_61 https://en.m.wikisource.org/wiki/The_Count_of_Monte_Cristo/C...

38 points by Clepsydra 8y ago

One of the Discworld novels uses 'hacking' of multi-shutter semaphore towers as a plot point, by way of a gibberish sequence that can physically destroy the poorly engineered towers by slamming the shutters around at a resonant frequency (book) or cause the controls to physically seize up in a jammed position (film).

smallnamespace 8y ago
Real physical layers are quite vulnerable to this type of disruption and we actually mitigate this by paying rather high costs at the bit encoding level: https://en.wikipedia.org/wiki/8b/10b_encoding https://en.wikipedia.org/wiki/8b/10b_encoding https://en.wikipedia.org/wiki/64b/66b_encoding https://en.wikipedia.org/wiki/64b/66b_encoding https://en.wikipedia.org/wiki/Scrambler https://en.wikipedia.org/wiki/Scrambler
sophacles 8y ago
GNU Terry Pratchett http://www.gnuterrypratchett.com/ http://www.gnuterrypratchett.com/

paxy 8y ago

I'm pretty sure you can go back a lot further if you keep redefining the meaning of the word. Would some Roman general sending a spy into an opposing camp to steal codes be a cyber attack?

tzar 8y ago
The semaphore ploy strikes me as much more rooted in information, thus the description “cyber”. Stealing code is just theft, since the object of theft doesn't generally qualify the act (though it does in some cases).
lainga 8y ago
It is cyber in the sense of cybernetics ("the mathematical study of communication and control in the animal and the machine")
- kazinator 8y ago
  Let's see: https://www.etymonline.com/word/cybernetic#etymonline_v_36576 https://www.etymonline.com/word/cybernetic#etymonline_v_3657... "kybernetikos": Greek for "good at steering". Thus military attack involving an excellent steering maneuver is cyber attack.
  - Udik 8y ago
    Not too sure about the -tikos thing and the "good at" part, google (and the Oxford dictionary) gives etymology from kubernetes, pilot. Cybernetics is then, I guess, the science of (automatic) control, so nothing to do with "steering".
kazinator 8y ago
Does the Trojan Horse ring a bell?

zitterbewegung 8y ago

A better title would be the first hack of a financial market.

dang 8y ago
We've given the article a narrower title above, based on its actual content.

kazinator 8y ago

An inside job to borrow some network resources does not qualify as an attack on a network. This is like using your neighbor's Wi-Fi without permission being a "cyber attack".

zitterbewegung 8y ago
It is a government owned network that only allowed government traffic. It’s more like borrowing Siprnet...

jgeralnik 8y ago

My favorite story that I've talked about as the first "cyber" attack for years is from the time of the second Jewish temple, a few hundred years B.C. The Jewish calendar is based on the lunar cycle, and each month has either 29 or 30 days. Back then the length of months was not declared in advance. Instead, each month when a new moon was spotted witnesses could come to the religious court (called Sanhedrin) in Jerusalem claiming to have witnessed the new moon, and if the court accepted their testimony that day was declared the first of the month. To spread the word they would light a bonfire on a mountain, and every city that saw the bonfire would light their own and the message would quickly spread. (Similar to the semaphore system but only transmitting a single bit. Because the signal was just a fire at nighttime it would reach very far very quickly) Now, there was a small offshoot of Jews known as Karaits, whose interpretation of the Bible caused them to believe that passover must fall on a Sunday (the passage says something like "on the first day [of passover] " which can also be interpeted as "on Sunday"). "Mainstream" (at the time) Judaism did not agree with this interpretation. So what did the Karaits do? On the day that they wanted to be the first of the month (such that Passover, the fourteenth of the month, would fall on Sunday) they went to a nearby mountain and lit a bonfire. The other cities saw the bonfire and assumed it was legitimate, and so lit their own bonfires as well, and the fake message was spread. Classic man in the middle attack! So how did the Sanhedrin solve this issue? One way would be to sign their messages with RSA. Generate two large prime numbers p and q, have N=p*q and e=65537 be public, known parameters, figure out how to properly pad your messages so that you're not using a completely broken system, and sign the hash of the message with phi=(p-1)(q-1). There are a few problems with this. First, it's difficult to transmit 1024 bit messages over fire signals. Second - Rivest, Shamir and Adelman wouldn't be born for another 2000 years or so. Instead, the solution was to move to a different, more secure medium. Instead of lighting a fire, the Sanhedrin would send messengers to spread the word to nearby cities, and each city would send their own messengers to their neighbors. Each city would trust their immediate neighbors and a chain of trust could be established. The problem with the new solution is that it is much much slower than the traditional fire signals, so some far off places wouldn't know when the previous month started by the time the 14th of the month (passover) came around. To handle this, far off cities would celebrate two days of passover - one on the 14th assuming the previous month had 29 days, and one on the following day under the assumption that there were 30 days. To this day, the tradition of Jews in far off places celebrating 2 days of holidays remains. Orthodox religious Jews outside of Israel celebrate 2 days of each of the 3 major Jewish holidays out of a sense of "doubt" over which day the holiday actually falls on because the message would not arrive in time.

twiceaday 8y ago

Tom Scott did a video on this: https://www.youtube.com/watch?v=cPeVsniB7b0 https://www.youtube.com/watch?v=cPeVsniB7b0

imglorp 8y ago

> Looking back at the incident now, one can see that the French semaphore system was hopelessly insecure,... Actually, this would have been an ideal application of one-time-pads for end-to-end encryption of the sensitive information, needing no modern tech beyond pencil and paper. The pads could have been passed by hand between participants and would have been highly secure.

Udik 8y ago
It seems that messages were actually "encrypted" and "compressed" with a shared code book. But encryption wouldn't have changed anything in this particular case: nobody was altering or decoding the authentic messages, they were just piggybacking on the error correction codes to insert more information. Very clever.

Udik 8y ago

Very interesting. A less clever but very similar attack, certainly inspired by this, is the described in the great classic "the Count of Monte Cristo", written ten years later. The Count manages to ruin one of his enemies by bribing a semaphore operator in transmitting news of a political unrest in Spain, causing a collapse in the value of foreign bonds (if I remember the details correctly).

credit_guy 8y ago

Most likely this event inspired Dumas, who put a similar hack in his Count of Monte Cristo novel. It’s funny that Dumas’s hack is less sophisticated that the one done in real life, but the whole chapter is quite well written. It’s available online at [1]. [1] https://en.m.wikisource.org/wiki/The_Count_of_Monte_Cristo/Chapter_61 https://en.m.wikisource.org/wiki/The_Count_of_Monte_Cristo/C...