uncrackable software/encryption algorithms

[Adrian]

do you think that this will EVER happen?

 

about a month ago on TechTV, they were explaining that quantum computers could crack ANY code within seconds/minutes. one example that they gave was the CIA's code...it would take months to years to crack it using todays computers, but only a matter of seconds with quantum computing...

 

for the past months, ive been tempted to write an encryption algorithms that cannot be cracked, but by todays standards, it would be semi-impossible...

 

 

what are your views on this?

[fafalone]

Just a matter of time... SSL encryption can be cracked, but it takes years on supercomputers. Quantum computers could do it significantly faster, but it will be another 20 years at least before we can build one sophisticated enough to be of any use (the most complex computation done to date with quantum computing is factoring 15 into 3 and 5)

The latest cryptography that cannot be cracked (with current technology) is quantum cryptography, because with existing levels of technology you cannot intercept the message without changing it. To get around that, we're talking 50+ years of quantum mechanics.

[aman]

Our physicists are starting to break the sub-quantum or microcosom barrier by examining high energy collisions of particles and seeing the effects. This barrier is where quantum cryptography operates and yes it cannot be examined without interfering with the message at this time. I think soon we will be able to understand on this lower level what exactly we are looking at and with sub-quantum small energy amounts be able to touch the information and not effect it. But then we would be able to develop sub-quantum real time codes that won't be able to be broken for a while. It's just my opinion but I think in our lifetimes we will know what is smaller than a quark and faster than C. I was blessed to be around some very intelligent, driven, people. There's a lot more of you out there.

Did you Know that a lot of the early codes in the fifties was wired into glass and plugged in on each flying mission so it could easily be destroyed and never reassembled. We've come a long ways

Just aman

[Radical Edward]

with current theory it is actually impossible to interfere with quantum encryption, since the wavefuntion is altered on a measurement. the actual description of how quenum encryption works is understandable, but quite lengthly, and diagrams make it easier. maybe I will have a go later if people are still interested.

 

anyway, ultimately the quantum element of the engryption isn't encryption at all, but rather the generation of a key, that cannot be seen by an outside observer. It has been mathematically proved that if a random key is the same size as the original message, then that message is immune to any attempts at deciphering it (not only is it mathematically provable, but it's also rather intuitive)

 

the way current encryption systems work is by the use of an asynchronous encryption key (the public key-private key system) rather than the more standard substitution ciphers. the maths to encrypt a message one way is usually simple (the public key) but to decrypt it, requires extra knowledge about the public key that was handed out.

 

there is a great book on encryption by Simon Singh called 'The Code Book' which has a good description of the development of encryption throughout the ages, from the basic substitution ciphers, through the enigma machine, and onto PGP and quantum encryption.

[Scorpio]

Anything that is decryptable is crackable. It is simple as that there is no such thing as an uncrackable encryption. No matter how complex anything is encrypted, if it can be decrypted it can be cracked.

[fafalone]

Technically yes, but if you use a key with 1 billion characters, which translates to (with a 255 character set) well.. it overflowed my TI-89 so it's well over 1x10^999 combinations, which would take billions of years to crack.

[Scorpio]

yea and just as long to decrypt. It would take a long time. and which makes it about as useless as tits on a bull. And still it is crackable.

[aman]

Is the only true advantage to this code the fact that the receiver will know it has been tapped into? I'll admit it has a compicated key but anything will be able to be decripted someday and maybe soon. Maybe by a large mass of cloned brain tissue.

Just aman

[Radical Edward]

Originally posted by Scorpio

there is no such thing as an uncrackable encryption

 

wrong, if you have a random key the same size as the message, encrypt with this key, and provided that no one gets hold of the key, the encryption is uncrackable.

 

there is also mathematical proof of this.

[Radical Edward]

Originally posted by aman

Is the only true advantage to this code the fact that the receiver will know it has been tapped into? I'll admit it has a compicated key but anything will be able to be decripted someday and maybe soon. Maybe by a large mass of cloned brain tissue.

Just aman

 

well you can figure out the likelihood of someone having attempted to intercept your key - it becomes quite easy to see of someone has looked at the key. again, once you have a key you are certain no-one has looked at that is the same size as your message, then it is uncrackable. this is quite easy to see: take all the values in the top row, and add them to the values in the bottom row

 

1 0 0 1 1 1 0 0 0 1 0 1 1 0 0 0 0 1 1 1 0 1 0

0 0 1 1 0 1 1 1 1 1 0 1 0 1 0 0 0 1 0 1 1 0 0

 

1 0 1 0 1 0 1 1 1 0 0 1 1 1 0 0 0 1 1 0 0 1 0

 

(1 + 1 = 0)

 

now if the top row is the key, and I destroy the key, you will never be able to find out what the middle row was, given the answer in the bottom row. granted there may be a certain non randomness to I have generated the key (this was one of the things that helped them decrypt the enigma code) but if it was generated by quantum means, it would be entirely random. (given our current understanding of quantum mechanics at the very least)

[Scorpio]

then it is undecryptable if it is undecryptable then there is no point of encryption then you are no longer encrypting you are corrupting.

[chris]

i think scorp is right. If you can bring it down to the simplest form (decrypting) it can be cracked. No matter how many possibilites (like the example 1x10^999) even if it takes billions of trillions of years to be cracked... can it be cracked? the answer, is yes. So it may be as useful as tits on a bull, but, arent there people into beastality....

[Radical Edward]

Originally posted by Scorpio

then it is undecryptable if it is undecryptable then there is no point of encryption then you are no longer encrypting you are corrupting.

 

eh? but if I give you the key, you can decrypt it. If I don't, it is impossible for you to decrypt it. I am not corrupting it at all.

 

remember that the person who recieves the message has to have a key. or is one of us missing something?

[chris]

ahhh. yes. very true. :cool2:

[John]

It is my belief that any encryption algorithm we create will ultimately be crackable. The human mind can work a way to crack any technology it creates to prevent cracking.

 

As for it being proven mathematically impossible, you're limiting yourself to our current understanding of mathematics. A while ago there was no 0, and/or there were no negative numbers, and/or there was/were no [insert most of the other stuff in our mathematical system]. Mathematics is evolving all the time, and using our current understanding of it to say something is certainly impossible isn't valid, in my opinion.

 

For the physics aspect, much the same applies, as was briefly mentioned in Radical Edward's earlier post.

[Radical Edward]

no.

 

It's been proved as in proved with self consistent maths, and my little thought experiment up there.

 

If you have a completely random key the same size as your initial message, there is no logical (hence mathematical) wat to decrypt it, unless you have the key. think about it, you'll see what I mean.

[John]

You're still limiting yourself to our current understanding of the universe. We've done things that were considered impossible by all logic a relatively short while ago, and cracking an encryption that seems uncrackable now will fall into that category, I think.

[Radical Edward]

trust me, the only way of decrypting the message is getting hold of the original key. If you can't get hold of the key, the message is completely random numbers added to non random numbers - which is still a random number. lets go for a stupidly primitive example

 

I want to send a number, but I don't want someone else to know the number. I pick a random key, which somehow I get to you (this is the tricky part) ... lets say this key is the number 4. only you and I know this. now I add it to the message, the number 8. you get 12. now lets say the spy gets hold of the transmission of the number 12, and the fact that I have added the key to the message. there is no way for him to know which of the numbers from 1-11 is the key, and which is the message.

 

 

If he can't get hold of the key, the message is uncrackable.

[fafalone]

No key based system is uncrackable, as all possible keys could be tried.

[fafalone]

and since most encrypted material is more than a few sentences, logical filters could easily detect whether the key was right.

[Radical Edward]

no, you would end up with every permutation of characters.

 

if I encrypted the message

 

hello there my name is edward

 

it is a string of 29 characters. put it in binary, and that is however many binary digits.

 

alternatively encrypt the message

 

plant the bomb at the embassy

 

it's also 29 characters long, and would be described with the same number of binary bits. without knowing the random key applied, it would be impossible to discern which message was right, even with a logical filter, and that is just two of the 27^29 of possible combinations of twenty nine characters (letters and speces only, no numbers, no punctuation) that makes sense.

[aman]

The two strings of binary digits would be different with the same key even though they each contain the same amount of information. At speeds and crunching power the random key is still decipherable theoretically.:flame:

Just aman

[Radical Edward]

in that case yes, but that's because I have used the key twice - I now effectively have a key that is half as long as the message. the longer the message is relative to the key, then the easier it is to dectypt. The more I use the same key, the easier it is to decrypt. If I have a new random key every time I send a message, that is the same size as the original key, then you can't decrypt it.

 

 

a simple exeriment is this. say my encryption sustem involves a number, in which I add that many letters to each letter of the message. and work sequentioally through the message, repeating the numbers as nescessary:

 

key = 3, letter = h, encrypted = k

 

now from that basis, you can't figure out what letter I started with. if you don't know the key. however If I foolishly use the same key five times, then you get

 

khoor

 

and it becomes more possible to figure out that I am saying 'hello' granted it is still a bit tricky with such a short message, but If I did it on a sentence, then it would become far easier, since you could use probability and knowledge of the English language to figure out what I was saying. alternatively, If I had a key that was 5 numbers long, all you could say is that I sent a message that was 5 letters long.

 

as i said in my first post:

 

It has been mathematically proved that if a random key is the same size as the original message, then that message is immune to any attempts at deciphering it

 

 

[edited to remove some ambiguity]

[aman]

That does make it difficult, for right now. We still have the option of the microcosom where the rules of math don't seem to apply. I see us going smaller and getting smarter but it's still a long ways off.

For now the system seems secure.

Just aman

[fafalone]

http://fafalone.hypermart.net/aol.html

 

Kind of related...

 

(and i tried posting this as an article here, but since for some reason vbulletin kills anything thats not basic html it wouldn't render right, and would have taken hours to convert from mso xml/dhtml to normal, so i gave up, apparently blike did too