Cryptanalysis - how widespread is this use of cryptography?

[cronxeh]

I was wondering about this for some time now. Considering certain applications in mathematics, particularly in cryptanalysis - how widespread is this use of cryptography? If I wanted to get into that field and had a degree in Math - how likely are foreign governments to use cryptography that the NSA would actually have to seek out mathematicians to work on them?

I guess what I'm asking is - how widespread is use of ciphers in telecommunications and how important is it to crack the cipher for certain government entitites if I wanted to get involved, and by that I mean devote a good chunk of my young adult years, in the field?

 

[Hurkyl]

The National Security Agency/Central Security Service is America’s cryptologic organization. It coordinates, directs, and performs highly specialized activities to protect U.S. information systems and produce foreign intelligence information.

...

NSA employs the country's premier cryptologists. It is said to be the largest employer of mathematicians in the United States and perhaps the world.

http://www.nsa.gov/about/index.cfm

 

[cronxeh]

yes yet somehow that has absolutely nothing to do with my question

 

[neurocomp2003]

Think about it from a banks perspective...if someone can crack NSA standard methods doesn't that mean banks are useless? so there is a continuous need to employ mathematicians/cs to attempt to see if these methods can be cracked(since technology enhances daily) and the day they are that group/person will be the richest in the world...
either from governments themselves, terrorists or they can just steal the money themselves...and when dealing with terrorist acts or governmental espionage it is vital to keep up with modern day cryptographic methods.

Two books you might find handy:
Koblitz-Cryptography and number theory
Cryptography for C++ i think it is. I just have the ebook

 

[cronxeh]

Well I do believe it would be possible to easily crack many currently accepted 'uncrackable' algorithms in next dozen years, through special cases in Quantum Physics, but this is more of a philosophical question - would foreign government actually be using cryptography for their purposes or would they simply crack under pressure and switch to some sort of a new method where they won't have a need for secrets

Although now that I think about the part about the internet and banks it does make sense for encryption

 

[Zurtex]

In the U.K as a mathematician who wants to get involved in that sort of work with government all you have to do is apply to GCHQ.

 

[shmoe]

Well I do believe it would be possible to easily crack many currently accepted 'uncrackable' algorithms in next dozen years,

Sure. Most cryptosystems aren't designed to give an everlasting security. The goal is to make your information secure for as long as it is sensitive.

would foreign government actually be using cryptography for their purposes or would they simply crack under pressure and switch to some sort of a new method where they won't have a need for secrets

What pressure? By the NSA trying to break theit codes? People will always have secrets, you can be sure they won't just give up because the nsa might be watching. I have no idea how much of nsa is devoted to prying into non-American nations, probably no one outside the nsa does either.

 

[exequor]

I was wondering; let's say you had someone develop a intricate encryption algorithm, you have it placed on some specialized encryptor chip then you kill the person who developed the algorithm and destroyed all records associated with it. If the chip was made to produce faulty data if there was an attempted hack. Do you think that this could be a foolproof encryption (although I don't believe that there could be such a thing)?

 

[cronxeh]

reverse engineering is the first thing they teach you in any intro to engineering class. you can be sure companies do that everyday on competitors' products, and anything humans make can be made and destroyed by other humans, its simply a matter of time

 

[neurocomp2003]

"Halting problem"... if there is a proof that it can't be hacked then sure...but like any of todays encoders...its the computational power we don't have..