ELINT / Signals Intelligence

[bash]

Originally Posted by swaghauler

The initial systems used 4-bit, 6-bit, and 8-bit (the most common in the Cold War era) representations for each LETTER in the LINK system. This meant that a 100-letter phrase would need 800 characters to be decrypted and placed in a proper order to read that phrase. Modern systems now use 128-bit encryption PER LETTER!

This is not quite accurate. A bit means "binary digit". Each bit can represent two states, on or off (1 or 0). A 100 letter phrases encoded as bytes (8 bits per byte) will weigh in at 800 bits, not characters.

Encryption like SAVILLE used is what's called a stream cipher. With a stream cipher a key fed into an algorithm to generate what's called a "key stream". Every bit of input data is combined with a bit of the key stream, usually with an exclusive-OR operation, to get an enciphered bit. A key stream essentially looks like random noise, least it should look like noise, and so long as you feed the same key into that algorithm if you feed in the enciphered bit you'll get the plaintext bit back out.

The size of the key is really describing the periodicity of the key stream. If the key was small, say only 8 bits, you could easily generate all the key streams from every possible key since there's only 256. Assuming your encryption algorithm doesn't have some other mathematical, process, or equipment weakness your key size increases the difficulty of someone trying every possible key. At 128 bits there's more possible keys that atoms in the universe IIRC. So it's not every letter being encoded with 128 bits but some pseudorandom extremely long pattern generated by the 128 bit key.

Originally Posted by swaghauler

Thus you will need a computer and special software that can do MILLIONS of operations per second to decrypt Digital Encryption. Therefore this is an IMPOSSIBLE Task to perform.

If a message is digitally encrypted I would put the difficulty as literally impossible. Unless there's some sort of key a character can get ahold of there's no practical way to crack the encryption. For military gear that means having the encrypted signals and having an intact key loading device and the appropriate encryption equipment. Without all that you'd need a billion years to find the encryption key.

[swaghauler]

Originally Posted by bash

This is not quite accurate. A bit means "binary digit". Each bit can represent two states, on or off (1 or 0). A 100 letter phrases encoded as bytes (8 bits per byte) will weigh in at 800 bits, not characters.

Encryption like SAVILLE used is what's called a stream cipher. With a stream cipher a key fed into an algorithm to generate what's called a "key stream". Every bit of input data is combined with a bit of the key stream, usually with an exclusive-OR operation, to get an enciphered bit. A key stream essentially looks like random noise, least it should look like noise, and so long as you feed the same key into that algorithm if you feed in the enciphered bit you'll get the plaintext bit back out.

The size of the key is really describing the periodicity of the key stream. If the key was small, say only 8 bits, you could easily generate all the key streams from every possible key since there's only 256. Assuming your encryption algorithm doesn't have some other mathematical, process, or equipment weakness your key size increases the difficulty of someone trying every possible key. At 128 bits there's more possible keys that atoms in the universe IIRC. So it's not every letter being encoded with 128 bits but some pseudorandom extremely long pattern generated by the 128 bit key.



If a message is digitally encrypted I would put the difficulty as literally impossible. Unless there's some sort of key a character can get ahold of there's no practical way to crack the encryption. For military gear that means having the encrypted signals and having an intact key loading device and the appropriate encryption equipment. Without all that you'd need a billion years to find the encryption key.

You undoubtedly have more experience than I do, so I'll take your word for it.

I too require PCs to have a powerful computer with specialized decryption software to have an IMPOSSIBLE chance to break digital encryption. I note that in my previous post.

[bash]

Originally Posted by swaghauler

You undoubtedly have more experience than I do, so I'll take your word for it.

I too require PCs to have a powerful computer with specialized decryption software to have an IMPOSSIBLE chance to break digital encryption. I note that in my previous post.

A meant literally impossible as in can't accomplish rather than simply Impossible skill check difficulty.

[pmulcahy11b]

Originally Posted by bash

A meant literally impossible as in can't accomplish rather than simply Impossible skill check difficulty.

I have difficulty believing that. One thing we've all learned in the modern computing age is that anything can be hacked.

[bash]

Originally Posted by pmulcahy11b

I have difficulty believing that. One thing we've all learned in the modern computing age is that anything can be hacked.

Hacked is a highly qualified statement. There's lots of components of communication systems that can be "hacked". Some parts are much easier than others, the encryption systems tends to sit more on the computationally infeasible end of the spectrum. Most successful hacks are more on the user side of a system like brute force guessing poorly generated/reused encryption keys.

So I'm definitely not saying systems are unhackable, it's just the PCs aren't going to be able to "hack" encrypted comms with a pre-TDM laptop they might have with them. Even if they were hauling around a supercomputer they couldn't expect to brute force any military encryption.

[Homer]

I think comms using reprogrammable encryption are going to be infrequent almost three years post tdm. It’d be easier to produce a paper SOI that gets couriered out and changed on a schedule than go to That’s essentially an encryption guide or code book (brevity codes) similar to JN25 or Admiralty Code. With time, an SOI or code can be compromised, mainly due to induced human error or capture. That countermeasure to that is introducing a new edition frequently. Couriers could see more use, with light aircraft and the remaining helicopters falling back into a liaison and courier role.

That will be another challenge for PCs as they re-enter friendly lines without current challenge/password or recognition signals, or counter “friendlies” with different signals. There’s adventure seeds to capture an SOI or codebook, interdict a courier, or protect the same. High priority missions may see PCs issued with encrypted comms gear and COMSEC equipment (ANCD, KYK, etc).

Originally Posted by Homer

That’s essentially an encryption guide or code book (brevity codes) similar to JN25 or Admiralty Code.

Speaking of brevity codes:

ZBM2

[Desert Mariner]

Originally Posted by bash

Hacked is a highly qualified statement. There's lots of components of communication systems that can be "hacked". Some parts are much easier than others, the encryption systems tends to sit more on the computationally infeasible end of the spectrum. Most successful hacks are more on the user side of a system like brute force guessing poorly generated/reused encryption keys.

So I'm definitely not saying systems are unhackable, it's just the PCs aren't going to be able to "hack" encrypted comms with a pre-TDM laptop they might have with them. Even if they were hauling around a supercomputer they couldn't expect to brute force any military encryption.

Some examples of the use of brute force, from RSA Labs Encryption Challenges, using distributed networks, not a stand-alone machine.

• 56-bit cracked in 250 days by 16,738 total participants
• 64-bit cracked in 1757 days (4.8 years) by 327,856 participants
• 72-bit remains uncracked after 7,241 days (19.8 years) and 143,497 participants (projected time remaining 27,828 days (76+ years))

To me, the question isn’t, Can you crack the encryption?, but Is it worth doing so? In the case of the above 72-bit key, even if you manage to crack it, does a single decrypt (or even a whole day’s worth of decrypts) provide any usable intelligence decades after the fact?

[bobcat]

for more modern setting two pieces of SIGINT gear i love to implement in more modern settings. the Wifi Pineapple and the Dope Scope. they balance each other out because the wifi pineapple is mostly intended for infiltrating networks (it can also be used as a wifi extender), and the dopescope is built to find anything emitting wifi signals. granted your range may vary and most the time i just put both at 50M and then adjust it based on terrain.

of course even without purpose built systems, with access to cheap SDR's, beer cans, and duct tape you can put together some surprisingly effecting radio direction finding equipment. might not be ideal for cracking crypto but you can do surprising things with very little.

[bash]

Originally Posted by Desert Mariner

Some examples of the use of brute force, from RSA Labs Encryption Challenges, using distributed networks, not a stand-alone machine.

• 56-bit cracked in 250 days by 16,738 total participants
• 64-bit cracked in 1757 days (4.8 years) by 327,856 participants
• 72-bit remains uncracked after 7,241 days (19.8 years) and 143,497 participants (projected time remaining 27,828 days (76+ years))

To me, the question isn’t, Can you crack the encryption?, but Is it worth doing so? In the case of the above 72-bit key, even if you manage to crack it, does a single decrypt (or even a whole day’s worth of decrypts) provide any usable intelligence decades after the fact?

Note that the SAVILLE system uses a 120-bit key. It would take trillions of years to brute force a single key if you could build a system that tried a quadrillion keys per second. So the time taken to recover a single key with a brute force search (even with mathematical principals like the Birthday Paradox) is directly related to a functionality of decryption. If a single key takes a trillion years to recover the practicality of the system is effectively zero.