PDA

View Full Version : 'Uncrackable' codes


PingDit
9th Sep 2013, 22:29
Very interesting, with some serious applications in the future...

A system that allows electronic messages to be sent with complete secrecy could be on the verge of expanding beyond niche applications.

BBC News - 'Uncrackable' codes set for step up (http://www.bbc.co.uk/news/science-environment-23946488)

Dak Man
9th Sep 2013, 22:39
Probably has some back door left ajar for the snoopers to have a look see.

Airborne Aircrew
9th Sep 2013, 22:51
Dak has it... :D:D:D

West Coast
9th Sep 2013, 22:59
Some pencil protector wearing nerd at the NSA is rolling up his sleeves looking forward to a new challenge.

I say jokingly, but it would certainly be a target of them.

500N
9th Sep 2013, 23:25
I see no one has cracked the Kryptos code at the CIA yet.

PingDit
9th Sep 2013, 23:44
Indeed WC. Toward the end of the article:

"The condition of quantum cryptography relies on certain rules that need to be obeyed - only then is it unconditionally safe. The newly proposed protocol is 'breakable' by middlemen attacks."

So even the critics are suggesting that if the rules ARE followed, it could be unconditionally safe...
That'd be a first!

West Coast
10th Sep 2013, 01:15
500

Are you sure?

Documents Reveal How the NSA Cracked the Kryptos Sculpture Years Before the CIA | Threat Level | Wired.com (http://www.wired.com/threatlevel/2013/07/nsa-cracked-kryptos-before-cia/)

500N
10th Sep 2013, 01:20
West Coast

Sorry, I meant the fourth message.

I knew the first three had been cracked

West Coast
10th Sep 2013, 01:31
It couldn't sit well in the CIA that the NSA scooped them on the first three.

Groundgripper
10th Sep 2013, 08:35
No doubt they'll call it the Titanic Code:E

GG

OFSO
10th Sep 2013, 10:49
Codes are solved by cracking repetition.

Simple tear-off books with lists of dedicated four-letter codes which are used once and thrown away, where the sender and the recipient have the only books, cannot be cracked.

One day DUPS may mean "get the kettle on", the next ORTY may mean "have you fed the cat ?" Both may be reused for totally diferent meanings.

Anyone know what I'm saying here ? No, you can't.

GRTD
OPDR
NMNP
WSER

rgbrock1
10th Sep 2013, 13:24
ALL codes can be cracked eventually. The only code which cannot ever be cracked is a one-time pad. One-time pads have their own inherent problems in implemention but they cannot be cracked if implemented correctly.

TWT
10th Sep 2013, 13:36
Didn't OFSO just describe a one time pad ?

rgbrock1
10th Sep 2013, 13:45
What OFSO described is not really a one-time pad. Similar but not exactly. What he described is a one-time cipher but not a pad. One-time pads use random keys which result in a ciphertext.

er340790
10th Sep 2013, 13:51
All sounds very clever...

HOWEVER, I suspect Alan Turing may have hit the nail on the head... any code set by mechanical, computational or other means devised by a human CAN be broken. You just need a tad more of the same power than it took to devise it...

And :mad: time, of course. :}

TWT
10th Sep 2013, 13:53
Aha ! Thanks for clearing that up that RG.

I read a book about the history of encryption a while ago.The earliest scheme known was in ancient Egypt.A slave would have his head shaved and the message would be written in indelible lettering on his noggin.After a while,his hair would grow back,covering the message.Off he would go on his journey and at the other end,the recipient would simply decrypt the message with a razor.

Of course,only non-urgent messages were sent by this method :O

A 'one-time slave' :p

mmmmmmmmmmmmmmmm
mmmmmmmmmmmmmmmmm
mmmmmmmmmmmmm
mmmmmmmmmmmmmmmmm

OFSO
10th Sep 2013, 18:55
Ask your wife (or any woman) to describe a location to you (or any man) over the phone.

Completely uncrackable.

rgbrock1
10th Sep 2013, 19:01
OFSO:

Or, in my case, I can phone my MiL (banish that thought) and ask her any question which will result in a stream of answers (babbling) intelligible only to those existing in an alternate universe.

She's capable of bring the NSA, or any other three-letter US agency, to its knees begging for mercy and deliverance. :ok:

racedo
10th Sep 2013, 19:04
NSA and PRISM is not really that good, reading something that someone is sent is wow so exciting...............not.

Bunch of clowns will really get people believing in them if they ever figure out what a woman is thinking.

Capot
10th Sep 2013, 19:44
I used a 1-time pad for 3 years running round the desert with a cloth on my head; sent with key using a 19 set and long dipole operated by an Arab signaller with speeds matching Marconi's best.

As has been mentioned, a 1-time pad, used properly, is unbreakable, and many attempts have been carried out with huge computing resources.

I'm sure someone here will know the arithmetic involved, but I don't.

By the way, the 1 time pad used groups of 5 letters. Figures were not used in the encrypted text.

ricardian
10th Sep 2013, 20:03
The VENONA project broke OTP but only because the OTP was not implemented correctly. It did take a long, long time!

Loose rivets
10th Sep 2013, 20:27
Richard Phillips has a lot to say about this in his first book of the Rho Agenda series. In The Second Ship, data has to be sent to the NSA without them knowing where it came from. Some very bright youngsters come up with a plan.

I was . . . am, very worried about the hyper-ramble I go into getting my concept of a 5th force established in the minds of my readers. However, in Phillips' best-selling books, he really goes to town on the quantum physics of the above NSA issue. It requires a reader with some dedication to understand the story, but now, just imagine for a moment, if it wasn't fiction how hard would it be to fool the most highly funded code-breakers in the world?

radeng
10th Sep 2013, 20:36
With enough computing power, all possible combinations in a OTP could be tried until something came out in the appropriate language that made sense. However, there could be many possible 'sensible' solutions, and choosing the right one might be difficult - especially if something like "Auntie Maude is coming to tea on Tuesday" is the right decode but means something important...

Teldorserious
11th Sep 2013, 05:17
I am pretty sure the sum total of cipher knowledge in this forum adds up to hitting the enter code on a software program to make a private key.

All the software has had backdoors built in, hence rendering your knowledge of encryption useless. It's like relying on the autopilot, hard for you to renounce your religion, it's all you know.

MG23
11th Sep 2013, 05:32
However, there could be many possible 'sensible' solutions, and choosing the right one might be difficult - especially if something like "Auntie Maude is coming to tea on Tuesday" is the right decode but means something important...

That's rather the point: a message encrypted correctly with a one-time pad can decode into any message of the same size. There is no way to know which of those messages was actually being sent, if you don't have the pad.

cattletruck
11th Sep 2013, 12:10
1-time pad can't encrypt say an image of the front line that is transmitted to the intelligence office.

1-time pad reminds me of how hieroglyphics were cracked with the discovery of the Rosetta Stone.

radeng
11th Sep 2013, 12:16
>1-time pad can't encrypt say an image of the front line that is transmitted to the intelligence office.<

Why not? The image is a string of 0s and 1s representing pixels: they can be encoded in various ways and the OTP made a memory that can be only read once. There may be a difficulty if the message gets corrupted, but a lot then depends on how much corruption there is and how good the error correction mechanism is.

cattletruck
11th Sep 2013, 12:35
I was speaking to a cannon operator some years ago and what impressed me most was the radio equipment he used to get the co-ordinates and feed back on his cannon's effectiveness. We discussed ciphers etc and to me it didn't sound like they used 1-time pad.

rgbrock1
11th Sep 2013, 12:51
cattletruck:

I was one of those canon "operators" you wrote about. However, it was many moons ago. (Canon operators are more affectionately known as gun bunnies or dumb ass gun bunnies. :})

I don't know how it is in the British army's artillery corp but back in the day comm gear/radios in the American artillery were rudimentary. And nothing on the gun/canon-side of things was encrypted. The encryption and all the high-tech radio equipment was borne by the Fire Direction Center (FDC) who obtained fire missions from the attached Forward Observers (FO's). The FO's gave FDC the grid coordinates of the target, via encrypted comms transmission) and then FDC took that information and computed the trajectory of the round to ensure there might be some "steel on the target."
These computations were then transmitted to the dumb ass gun bunnies and we did our thing and then pulled the lanyard.

FWIW, these are the guns I had the privilege of firing many times. (The M109A2 155mm self-propelled howitzer. Able to wreak havoc in a single bound!)

http://www.41afdva.net/3/Geschied08_M109A2-90.JPG

cattletruck
11th Sep 2013, 13:10
rgb, that looks like fun :E.

Yeah I was once 'interviewed', actually I was sussed out by a bloke working at the defence signals directorate who was looking for his replacement after his retirement. His job was to 'assign the ciphers' and ensure equipment was complying, with a bit of problem solving.

I must be a bunny because I failed the test by not realising I was actually being 'interviewed' at the time, though it was clear as daylight the following day. :ugh:

Gibon2
11th Sep 2013, 13:39
The one-time pad is currently the only known encryption algorithm that can be mathematically proved to be uncrackable (provided it is properly used, etc etc). As observed above, even if you use infinite computing power to try all possible combinations, this only gives you all possible messages of that length. So if you have encoded the 14-character message ATTACK AT DAWN with OTP, a code-breaker trying all combinations (and there are around 10^22 of them) will indeed find ATTACK AT DAWN, but will also find ATTACK AT DUSK, ATTACK AT 1300, ATTACK AT ONCE, DO NOT PROCEED, CANCEL ASSAULT, CEASE FIRE NOW, HOLD POSITIONS, ABANDON ATTACK, AIR STRIKE NOW, NO AIR SUPPORT, and literally millions of other plausible messages.

OTP can be implemented electronically and can encode anything that can be encoded by other methods. But despite being uncrackable, OTP is rarely used as it has some serious practical drawbacks. The two main problems are the need to securely exchange the key (which must be as long as the message - so if you can securely send the key, why not just send the message?) and the need for the key to be completely random and never re-used (truly random keys are not easy to generate in practice).

The most famous use of OTP is to encode the Washington-Moscow hotline, a unique application to which it is well suited for a number of obvious and more subtle reasons.

MagnusP
11th Sep 2013, 13:57
When my uncle was sent overseas in WW2, he used a one-time pad to let his parents know where he'd been sent (information which was removed by censors from letters home). Each member of the extended family was allocated to an arena of war and an enquiry as to how Auntie Jessie was keeping meant that he'd been sent to Egypt, for example.

radeng
11th Sep 2013, 15:47
A method used by behind the lines radio operators in Yugoslavia in WW2 was, before they left Cairo, to buy 2 copies of a paperback book, one of which was left at Cairo. This then became their code book, with the words they wanted referred to by a 5 letter code: that gave the page number and word number. Offered nearly 12 million possible locations for a word.

But rather slow....if relatively secure.

MG23
11th Sep 2013, 16:54
I was speaking to a cannon operator some years ago and what impressed me most was the radio equipment he used to get the co-ordinates and feed back on his cannon's effectiveness. We discussed ciphers etc and to me it didn't sound like they used 1-time pad.

For something like that, a one-time pad would be excessive. Doesn't matter if the bad guys can decipher yesterday's messages after the computer churns through them for a few hours, because they already know where you were firing at, since they were probably there and getting blown up at the time.

A relatively simple cipher with a key that's changed on a daily basis should be enough, both to keep the messages secret until they're too old to matter, and authenticate them to the people firing the cannon.

A cipher to protect your credit card data going to Amazon, on the other hand, has to keep it secure for a few years until the card expires.

MG23
11th Sep 2013, 17:02
I am pretty sure the sum total of cipher knowledge in this forum adds up to hitting the enter code on a software program to make a private key.

Then you'd be wrong :).

I can't vouch for all software, but I know the stuff I wrote years ago had no back doors, unless the actual encryption algorithms we used had back doors. No-one has ever found one.

Capot
11th Sep 2013, 22:11
A relatively simple cipher with a key that's changed on a daily basis should be enough, both to keep the messages secret until they're too old to matter, and authenticate them to the people firing the cannonSlidex, anyone?

Now I'm really showing my age.

Slidex was a battlefield message coding device that probably kept your intentions secret for 30 minutes with an unsophisticated enemy, or 3 days in Arabia. I can only vaguely remember how it worked, but I do recall a form of matrix with a vertical and horizontal key on the side, or something like that.

By the way, all UK artillery firing comms were in clear; there would have been little point in keeping them secret. But we dispensed with call signs and all that "over and out" stuff as soon as ranging shots began.

AtomKraft
11th Sep 2013, 22:32
We used MAPCO for low level grid ref encryption.

Can't see the point of encrypting artillery info.

As the splash is known, you'd just be helping the enemy to crack your code!

Isn't a 'true' one time pad just a book- held by all parties involved?

Simple, and clearly impossible to crack.

Does anyone else remember the 'Radio Moscow' transmissions which consisted entirely of 5 (I think) figure groups in morse? It was transmitted 24/7.

Sounded quite spooky to listen to as there were clearly folk furtively scribbling away somewhere....:uhoh:

MG23
11th Sep 2013, 22:35
Slidex, anyone?

Interesting, I've never heard of that one before.

Google found some pictures, but still isn't entirely clear as to how it was used:

Slidex (http://www.jproc.ca/crypto/slidex.html)

ShyTorque
12th Sep 2013, 09:53
Just phone up instead and talk Scottish. No-one can understand that.

Capot
12th Sep 2013, 11:18
Isn't a 'true' one time pad just a book- held by all parties involved?Not exactly; it was a pad of tear off sheets, and the recipient had a matching pad. (Recipients? Maybe, can't recall. In our case it was always 1:1, the other '1' being Int. HQ.)

Each sheet was destroyed after use, hence the term "one-time pad".

I'm straining, and failing, to remember the procedure; it started with writing the message letters and numerals above the 5-letter groups in the tear-off pad, but what happened then is lost in the mists of time.

The message started with information enabling the recipient to use the correct decyphering pad/sheet, also "one-time".

A lost/stolen pad was worthless for code-breaking purposes. It was simply 50 - 100 (?) tear-off sheets of meaningless random 5-letter groups, which anyone could generate. None-the-less we did take extraordinary care not to lose them.

Re Slidex: but still isn't entirely clear as to how it was used:Reminded by that link....you set up the 2 "Key" strips with the code letters for the period (distributed by HQ) and then just read off the letters for a word like a map reference. So, if "Surrender Now!" had UR above it, and FU to its left, you would send/say "URFU", and the recipient would immediately take out his white handkerchief and wave it at the enemy.
I rather think that the numbers on the strips determined the position of the strip, and would be sent at the start of the message, to add a layer of complexity to it. Or something like that.

G-CPTN
12th Sep 2013, 11:55
Just phone up instead and talk Scottish. No-one can understand that.
I believe that 'Native Americans' were used for precisely this purpose.
(though not speaking Scottish of course)

TWT
12th Sep 2013, 13:58
Apparently,the Japanese never managed to crack the American ciphers in the Pacific in WW2.As mentioned,native American.Choctaw.The Japanese were expecting a cipher that decrypted to English.The Americans didn't have the same problems decrypting Japanese traffic.

ThreadBaron
12th Sep 2013, 16:39
On the subject, may I recommend Leo Marks', "Between Silk and Cyanide". The story of his 4 years, 1941-1945, as a codemaker and contains much technical information about codemaking.

The poem,

The life that I have
is all that I have.
And the life that I have
is yours.

The love that I have
of the life that I have
is yours and yours and yours.

A sleep I shall have.
A rest I shall have,
Yet death will be but a pause.

For the peace of my years
in the long green grass
will be yours and yours and yours.

was written by Marks as a code poem for Violette Szabo.