Notices
Results 1 to 31 of 31

Thread: Quantum computing and the future of cryptography

  1. #1 Quantum computing and the future of cryptography 
    Forum Junior
    Join Date
    Jul 2008
    Posts
    287
    Physics most probably allows for non-classical computation methods, which makes parallel computation of all possible inputs - like becoming more and more realistic quantum computers.
    I think we should widely discuss if it's a real threat and if yes - how to design cryptosystems resistant to such eventualities.
    Especially that we wouldn't know if someone would use it...
    Here is one discussion of this type:
    http://www.ecrypt.eu.org/stream/phor...?1,1021,page=1
    I wanted to start here one. Please share Your opinions and links to interesting discussions on this topic.

    There is know Shor's algorithm which breaks RSA. Generally cryptosystems have always a 'weakness' that make them prone to brute force attacks - that there is a key which properly decrypts the message. To make such attack we can use that there is practically only one key which makes that message decrypted with it has some significant correlation.
    So using a quantum computer we could use entangled all possible keys to decrypt the message, check for correlations and somehow extract some information about the correct key.
    This extraction is more difficult than it looks, but assuming that it's impossible could be dangerous.

    Quantum computers are extremely difficult to make and they most probably will be limited to relatively small number of qbits and time they can sustain entanglement - so I think that we can protect against imaginable QC by enforcing large amount of required computation. To make such cryptosystem practical, this computations should be done only once - specifically for given key. The next advantage of such preintialized cryptosystem, like based on asymmetric numeral systems, is that now the processing of the data can be extremely fast.

    Do You think it's a real threat?
    How strong it is - can be used only for some algebraic attacks or even for brute force?
    How to protect against such eventualities?
    What about public key cryptography???


    Reply With Quote  
     

  2.  
     

  3. #2 Re: Quantum computing and the future of cryptography 
    Suspended
    Join Date
    Apr 2008
    Posts
    2,178
    Quote Originally Posted by Jarek Duda
    Physics most probably allows for non-classical computation methods, which makes parallel computation of all possible inputs - like becoming more and more realistic quantum computers.
    I think we should widely discuss if it's a real threat and if yes - how to design cryptosystems resistant to such eventualities.
    Especially that we wouldn't know if someone would use it...
    Here is one discussion of this type:
    http://www.ecrypt.eu.org/stream/phor...?1,1021,page=1
    I wanted to start here one. Please share Your opinions and links to interesting discussions on this topic.

    There is know Shor's algorithm which breaks RSA. Generally cryptosystems have always a 'weakness' that make them prone to brute force attacks - that there is a key which properly decrypts the message. To make such attack we can use that there is practically only one key which makes that message decrypted with it has some significant correlation.
    So using a quantum computer we could use entangled all possible keys to decrypt the message, check for correlations and somehow extract some information about the correct key.
    This extraction is more difficult than it looks, but assuming that it's impossible could be dangerous.

    Quantum computers are extremely difficult to make and they most probably will be limited to relatively small number of qbits and time they can sustain entanglement - so I think that we can protect against imaginable QC by enforcing large amount of required computation. To make such cryptosystem practical, this computations should be done only once - specifically for given key. The next advantage of such preintialized cryptosystem, like based on asymmetric numeral systems, is that now the processing of the data can be extremely fast.

    Do You think it's a real threat?
    How strong it is - can be used only for some algebraic attacks or even for brute force?
    How to protect against such eventualities?
    What about public key cryptography???

    A computers strength is not that it thinks or imagines anything on its own. Or that it can do many things at once. A computers real strength, is that it can perform a trillion calculations on the same bit of information a trillion times, in minutes.

    A computes strength is its one track mind. Ha-ha.

    Encrypting is silly. Decrypting is even sillier. By the time you need to encrypt. The criminals are selling you the encrypting and decrypting tools. Ha-ha.


    Sincerely,


    William McCormick


    Reply With Quote  
     

  4. #3  
    Forum Junior
    Join Date
    Jul 2008
    Posts
    287
    Performing trillions of operations is faaaaar from being enough to break standard cryptosystem.
    But the question is if physics allow to make huge amount of them parallely.
    And on this 'silly' question depends world's safeness...

    Some people believe that the only real algorithmic advantage of QC is the Shor's algorithm. Maybe they are right, but ... QC can theoretically make all calculations at once (is almost nondeterministic Turing machine) and the only problem is with the extraction. I'll show how to enhance it, but it would be strange if basic QC wouldn't already allow for more, maybe even solving NP in polynomial time.
    Generally what physics do is solving its own partial differential equation - that means processing simultaneously infinitely many (or huge amount) degrees of freedom. I see a few ways of 'giving' there some NP-problem to be solved ... maybe some of them are practical ...

    In such case 1024 bit keys, which are more than atoms in the universe, would be a piece of cake - elongating key may be not enough. Especially that I believe that we could easily protect private key cryptosystems against such eventualities by using preinitialized ones.
    But designing public key cipher is much more difficult and I have no idea how to make a protected one???

    The next argument that we should take such scenarios seriously is that maybe basic QC is not the only possibility for massive parallel computation physics gives us. First of all there is so called Feynman-Stueckelberg effect
    http://groups.google.com/group/sci.p...10b4e5cbda1108
    which hasn't been taken seriously, but maybe it will change in a few months in LHC ... but such computer would require (huge?) accelerator.

    The other option can be (quantum) loop computers
    http://forums.devshed.com/security-a...ms-580926.html
    I'm strongly confused about this idea, especially for classical computers.
    But ... if used for quantum computation, such feedback should amplify the correct solution (wavefunction), making the rest of them vanish.
    It couldn't be standard approach to QC in which we use some sequence of for example external fields on some lattice of atoms.

    We would need a circuit which allows to sustain entanglement of many calculations.
    Observe that similarly to benzene, (-CH=CH-) sequence can be in quantum superposition with shifted one (=CH-CH=) - we could use such molecule as a wire for qbits. Unfortunately it has some resistance, but there are know such superconductors also.
    We know also transistors made of single molecule - they are irreversible so would destroy entanglement, but there should be possible also quantum gates made this way.
    The question is if such molecular quantum computers could sustain entanglement for practically long time ... There is also problem with auxiliary variables - we need a lot of them because in QC all calculations has to be reversible. They cannot be sent in the loop - they should be treated in some special way to not destroy the entanglement...
    ... but maybe ?

    Probably physics doesn't allow to solve NP in polynomial time, but I'm far from being sure of it.
    And I believe that preinitialized cryptosytems should be practically protected against all presented hypothetical possibilities. And this protection is achieved practically for free.

    -------------------------------------

    Ok - I was too pessimistic about public key cryptography - we should be able to make protected and practical hybrid systems - public key cipher for very short message like a key for a secret-key cipher or a hash value for authentication.

    Most generally, public key is a parameter of some transformation which is extremely difficult to reverse. But there is the private key - some kind of 'clues' which make this reverse easy.
    So if someone could solve quickly NP problems:
    - he could try all possible 'clues' and for example check if for some block encrypting and then decrypting gives the same block. If yes - he could try a few more different blocks to be sure it's the correct private key, but there is also more dangerous attack:
    - searching not for these 'clues' but straightforward for the reverse function: having encrypted message in a form of independent blocks, for each block he could try to encode all possible input blocks with the public key to get the given block.

    So to protect it in analogy to secret-key ciphers, we rather have to make that encoding already require extremely large amount of calculations. The problem is that this time these huge calculations cannot be just made while initialization like before, but has to be made for each block - it could be practically used only for extremely short messages, like the key for a secret-key cipher or a hash value.
    Reply With Quote  
     

  5. #4  
    Suspended
    Join Date
    Apr 2008
    Posts
    2,178
    Quote Originally Posted by Jarek Duda
    Performing trillions of operations is faaaaar from being enough to break standard cryptosystem.
    But the question is if physics allow to make huge amount of them parallely.
    And on this 'silly' question depends world's safeness...

    Some people believe that the only real algorithmic advantage of QC is the Shor's algorithm. Maybe they are right, but ... QC can theoretically make all calculations at once (is almost nondeterministic Turing machine) and the only problem is with the extraction. I'll show how to enhance it, but it would be strange if basic QC wouldn't already allow for more, maybe even solving NP in polynomial time.
    Generally what physics do is solving its own partial differential equation - that means processing simultaneously infinitely many (or huge amount) degrees of freedom. I see a few ways of 'giving' there some NP-problem to be solved ... maybe some of them are practical ...

    In such case 1024 bit keys, which are more than atoms in the universe, would be a piece of cake - elongating key may be not enough. Especially that I believe that we could easily protect private key cryptosystems against such eventualities by using preinitialized ones.
    But designing public key cipher is much more difficult and I have no idea how to make a protected one???

    The next argument that we should take such scenarios seriously is that maybe basic QC is not the only possibility for massive parallel computation physics gives us. First of all there is so called Feynman-Stueckelberg effect
    http://groups.google.com/group/sci.p...10b4e5cbda1108
    which hasn't been taken seriously, but maybe it will change in a few months in LHC ... but such computer would require (huge?) accelerator.

    The other option can be (quantum) loop computers
    http://forums.devshed.com/security-a...ms-580926.html
    I'm strongly confused about this idea, especially for classical computers.
    But ... if used for quantum computation, such feedback should amplify the correct solution (wavefunction), making the rest of them vanish.
    It couldn't be standard approach to QC in which we use some sequence of for example external fields on some lattice of atoms.

    We would need a circuit which allows to sustain entanglement of many calculations.
    Observe that similarly to benzene, (-CH=CH-) sequence can be in quantum superposition with shifted one (=CH-CH=) - we could use such molecule as a wire for qbits. Unfortunately it has some resistance, but there are know such superconductors also.
    We know also transistors made of single molecule - they are irreversible so would destroy entanglement, but there should be possible also quantum gates made this way.
    The question is if such molecular quantum computers could sustain entanglement for practically long time ... There is also problem with auxiliary variables - we need a lot of them because in QC all calculations has to be reversible. They cannot be sent in the loop - they should be treated in some special way to not destroy the entanglement...
    ... but maybe ?

    Probably physics doesn't allow to solve NP in polynomial time, but I'm far from being sure of it.
    And I believe that preinitialized cryptosytems should be practically protected against all presented hypothetical possibilities. And this protection is achieved practically for free.

    -------------------------------------

    Ok - I was too pessimistic about public key cryptography - we should be able to make protected and practical hybrid systems - public key cipher for very short message like a key for a secret-key cipher or a hash value for authentication.

    Most generally, public key is a parameter of some transformation which is extremely difficult to reverse. But there is the private key - some kind of 'clues' which make this reverse easy.
    So if someone could solve quickly NP problems:
    - he could try all possible 'clues' and for example check if for some block encrypting and then decrypting gives the same block. If yes - he could try a few more different blocks to be sure it's the correct private key, but there is also more dangerous attack:
    - searching not for these 'clues' but straightforward for the reverse function: having encrypted message in a form of independent blocks, for each block he could try to encode all possible input blocks with the public key to get the given block.

    So to protect it in analogy to secret-key ciphers, we rather have to make that encoding already require extremely large amount of calculations. The problem is that this time these huge calculations cannot be just made while initialization like before, but has to be made for each block - it could be practically used only for extremely short messages, like the key for a secret-key cipher or a hash value.
    No matter how many calculations this imaginary quantum computer can make, in some short period of time. Unlocking an encrypted message means that the message must contain some language that you know and understand. Or the computer will not uncover anything of value anyway. The computer has to be looking for something you told it was important or key information.

    Emotions can mask or give meaning to words and phrases. So it will always be a silly thing.

    We don't need quantum computers to decrypt. We just need the keys, as would any computer, or a man doing it manually.

    Military often changes their language so no one but someone with the keys can make sense of it, even if it is decrypted. However it is all fleeting. Because as soon as you recognize some pattern in destruction, you just isolate when keys where used.

    If you actually look at the history of counterintelligence. You will find it destroys the camaraderie of the men in your own army. Because only certain men are intrusted with certain keys. Encryption and secret languages, become a very self destructive communication block, to the military. I have heard men argue about what weapons were called at different times. They were in the same service, using the same weapon, but at different times.

    Then there is the phony communications, that the quantum computer cannot understand. You send millions of nearly coherent messages along with your real message, and it all becomes a guess. You give them millions of key words a second. To waste time on.

    To me it is all a waste of life.

    By the time you need decryption to win a war, your own country failed you. The enemy is right in your own capital.


    Sincerely,


    William McCormick
    Reply With Quote  
     

  6. #5  
    Forum Junior
    Join Date
    Jul 2008
    Posts
    287
    Any natural language (including mimicry, emotions...) You take will rather naturally have some correlations - only one key should decipher it into something with nontrivial correlations. This redundancy is kind of error correction method for our brains. We can reduce it greatly by using data compression - it's good protection against brute force attack, but is not perfect - we should assume that the intruder can know protocols and so while brute force he also makes decompression.

    But we are not talking only about natural language or military codes - kind of substitution ciphers. Assume somebody breaks cryptography used by banks for money transfers ... or used by missiles to communicate about the destination. It's quite imaginable scenario that he somehow know used protocols, so the safeness lies practically only in that he don't know codes.
    Checking all possible keys 'manually' would take maaaany billions of years ... but maybe physics allow to check all of them simultaneously ... like in QC...

    Of course there are many other ways safeness can be lost, but I believe this one can be well protected even in the age of nondeterministic Turing machines and not relying on that the intruder don't know protocols.

    Best,
    Jarek
    ps. here has started a discussion: http://groups.google.com/group/sci.c...2bd7fe1228afe#
    Reply With Quote  
     

  7. #6  
    Suspended
    Join Date
    Apr 2008
    Posts
    2,178
    Quote Originally Posted by Jarek Duda
    Any natural language (including mimics, emotions...) You take will rather naturally have some correlations - only one key should decipher it into something with nontrivial correlations. This redundancy is kind of error correction method for our brains. We can reduce it greatly by using data compression - it's good protection against brute force attack, but is not perfect - we should assume that the intruder can know protocols and so while brute force he also makes decompression.

    But we are not talking only about natural language or military codes - kind of substitution ciphers. Assume somebody breaks cryptography used by banks for money transfers ... or used by missiles to communicate about the destination. It's quite imaginable scenario that he somehow know used protocols, so the safeness lies practically only in that he don't know codes.
    Checking all possible keys 'manually' would take maaaany billions of years ... but maybe physics allow to check all of them simultaneously ... like in QC...

    Of course there are many other ways safeness can be lost, but I believe this one can be well protected even in the age of nondeterministic Turing machines and not relying on that the intruder don't know protocols.

    Best,
    Jarek
    ps. here has started a discussion: http://groups.google.com/group/sci.c...2bd7fe1228afe#
    The people that program for the banks are paid well not to give out the codes. Fire them or reduce their salaries. And they will show you a back door.

    I was a Network supervisor for a large chemical removal company, that handled Superfund sites for the EPA. Security was just to funny to me. We were playing Nsnipes over the fast bus loop network in the eighties. Because I created a system that turned hours of billing into fifteen minutes work.
    The guys in the back had the same clearance as the accountants. Ha-ha. Nobody else new what was happening. I could not even explain it to them.

    But the funniest was. This one fellow, thought I had not given him high enough clearance. So he brought in a professional programmer. To hack the system. Well, I used to hide the files on the boot disc, so the girls would not accidentally erase the files on the disc. Well this guy although probably a fine programmer, did not know you could do that. So he tried at it, and came to the conclusion that, I must have installed ROM boot chips. And gave up. When the owner heard, what happened he ended up firing him.

    The funniest thing was that, I even had hidden batch files so all you had to do was type your first name. And it would enter, "Login (Persons Name)" to save you key strokes and unneeded work. All the fellow had to do was type his name and his password and he could have accessed it like a supervisor. Which the owner originally wanted him to have those rights in case something happened to me.

    The defense plants or military, will give the codes to foreign countries when they feel a new system should be purchased. Often the other side already has them. By picking up a telephone and asking for them. Or by just watching us test. Or by their own projects successes and failures.

    All this stuff is to hide what we already know about the other side and our own leaders poor leadership. It is to hide our duty to our own country. It encrypts and complicates, our failure and lack of duty.

    There are other methods for breaking keys that are rather straight forward, and very fast. There is nothing you can do to protect information, if you know what the information is. And if you know what the information is. And you must. You really do not need much more, with all the tools available today.
    Lets face it though, you start moving money, the Secret Service will come a knocking. If you are not going to inflict chaos and destruction to hide your tracks. Best to leave it alone. That is what war is for.

    The bottom line is that when you have your best minds working on encryption and decryption, they miss the obvious and overlooked weaknesses. It is all a scam to hide your own failure from yourself and the "enemy".





    Sincerely,


    William McCormick
    Reply With Quote  
     

  8. #7  
    墨子 DaBOB's Avatar
    Join Date
    Jun 2006
    Posts
    1,674
    Wow, I know nothing about cryptology but I'm really enjoying this!
    Do not try and bend the spoon. That's impossible. Instead... only realize the truth. There is no spoon. Then you'll see that it is not the spoon that bends, it is only yourself. -Spoon Boy
    Reply With Quote  
     

  9. #8  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    Quantum cryptography. It isn't foolproof against someone listening in, but you'll know immediately if they do so you stop talking.
    Reply With Quote  
     

  10. #9  
    Suspended
    Join Date
    Apr 2008
    Posts
    2,178
    Quote Originally Posted by Numsgil
    Quantum cryptography. It isn't foolproof against someone listening in, but you'll know immediately if they do so you stop talking.
    I would not buy that, "cannot ease drop". If it is over air waves. You can always record the whole communication and then decrypt it later.

    If it is done over a hard system, you can always record it and decrypt it later.

    If you were going to ease drop on someone the best way would be to give them the secure equipment. It would be like American military buying chips from another country. There is no telling what is in them. By the time you chopped them up and examined them with an opti-scope you could make your own.

    Most people would not believe what you can do in a matter of hours with telephone numbers. You can go to any hard wire phone system and tap in, at any junction along the route back to the phone company. You can make it seem like you are a company or a person. From a telephone poll. Or from your car, or van nearby.

    I know you can create a second line through one of the more powerful companies in charge of the network. You can actually just call up and get a telephone number that was going to cable vision, switched to Verizon. So in just a few minutes the phone is ringing to a Verizon network. With just a phone call.

    It is a lot like web addresses. They can easily be redirected.

    With a computer network like cable vision, I believe I could install a battery powered router, or connect it to the low voltage lines, either at a house or business and duplicate any transaction, and send the information somewhere for decryption. Much test equipment is designed just for that purpose.

    Now the new FIOS although no system is totally safe. And I am sure someone somewhere can tell exactly what is being communicated over that network and record it. Is a pain to decrypt. Because you have to do a field splice of the optics. And you need the software and hardware that runs the Fios optic system. Or find a weak link between that system and the other system that you are taping.

    It could be done but it would be a more expensive thing to do.


    But I do not want to know anything anyone is saying. If I do I will ask them.



    Sincerely,


    William McCormick,
    Reply With Quote  
     

  11. #10  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    Geez Bill, at least read the article before you say it can't work...
    Reply With Quote  
     

  12. #11  
    Suspended
    Join Date
    Apr 2008
    Posts
    2,178
    Quote Originally Posted by Numsgil
    Geez Bill, at least read the article before you say it can't work...
    I did. There is no way, not to be able to record all the information being transmitted. There is no magic. If the antenna, tuner, receiver, or whatever device that is picking up a signal can pick it up. So can a hundred others. Complete for later decryption.

    Over the air, or over hard wire or optics. There is no place to hide. Ha-ha.


    I actually like it like that.

    It may require a directional antenna to knock out other signals, if you do not have control of a cell site. But it can certainly be done.

    But today they can probably isolate a cell site, and just record everything coming in for the duration they know someone is talking. And then later just decode it.


    Sincerely,


    William McCormick
    Reply With Quote  
     

  13. #12  
    Forum Junior
    Join Date
    Jul 2008
    Posts
    287
    Quantum cryptography is made that You cannot just intercept the signal - in standard approach You don't even know what to measure.
    But ... one can install herself in the middle - tell Alice that she is Bob, Bob that she is Alice ... and finally get the transmitted messages.

    It's only another protocol - we cannot rely on it.
    Especially that we should be able to do that encrypted message is just safe.
    Ok - there is always human factor which greatly reduces safeness. But in good protection after firing a person, it should be enough to change codes.
    Reply With Quote  
     

  14. #13  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    Quantum cryptography is entirely impossible for a 3rd party to listen in on, even with infinite resources, provided that:

    1. The attacker doesn't have physical access to the sender or receiver.
    2. Random numbers are generated from purely random means, like using quantum states.
    3. The sender and receiver both already know who each other is (that is, no impostors at the other end).

    source.

    This is also a good article on the matter.

    It can't be eavesdropped without either party knowing immediately. And it can't be cracked with a fast enough computer. You can't set up a listening device between the sender and listener and record then transmit the information, because to do so you have to guess at the state of the bit when you receive it. And you'll guess wrong, and the listener will be able to detect the errors and know something fishy is going on.
    Reply With Quote  
     

  15. #14  
    Forum Junior
    Join Date
    Jul 2008
    Posts
    287
    There is always protocol using which Bob finally receives the whole message.
    So if Eve install herself in the middle, say that she is Bob and use this protocol - she finally gets this message - doesn't she?
    Of course she can later use the same protocol to send it to Bob.
    Reply With Quote  
     

  16. #15  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    I'm not a crypto guru, so let me be sure I have the terms right. Alice is the sender, Bob is the listener, and Eve is the eavesdropper, right? Likewise my understanding of quantum cryptography is pretty limited, so take everything I say with a grain of salt. All the literature online is extremely convoluted.

    This is how I understand it to work: Alice and Bob need a secret key to encrypt their message. Let's say that their secret key is the same length as the message they want to send. Alice sends a single photon to Bob with a certain polarization (one of four states: 0, 45, 90, -45 degrees), chosen at random through quantum means (so it really is random). Bob receives the photon, and he has to guess which alignment it might be.

    "To send a key in Quantum Cryptography, simply send photons in one of four polarizations: -45, 0, 45, or 90 degrees. As you know, the receiver can measure, say, whether or not a photon is polarized at 90 degrees and if it is not then be sure than it was polarized at 0 degrees. Similarly the receiver can measure whether a photon was polarized at 45 degrees, and if it is not then it is surely polarized at -45 degrees. However the receiver can not measure both the 0 degree state and 45 degree state, since the first measurement destroys the information of the second one, regardless of which one is performed first." From this source.

    It's a 50/50 guess, so Bob hooks up the guessing to another quantum random number generator. If he guesses wrong he can't obtain any information about the polarization of the photon sent. If he guesses right he records the results, without revealing the results to anyone.

    So roughly half of the bits received by Bob were unintelligible, but Bob doesn't know this (the misses were recorded as random bits). Bob displays which way he measured the bits each time on a public channel. Alice compares this to the polarization of the photons she sent. She sends back a bit string telling him when he guessed right. The bits from when he guessed wrong are then discarded by Bob. Both Alice and Bob now have a shared secret key of some variable length. If Bob and Alice don't have enough bits, they repeat the process until they do.

    Now with their secret key, Alice bitwise xors the message with the key. The message can now be sent over public channels. Bob receives the message and xors it with the key, "decrypting" it back to the original message.

    Now let's say Eve listens in on the publicly sent encrypted message. She can't decrypt it using brute force methods, because every possible key will, when Xor'ed to the encrypted message, result in every possible original message. The orders to invade Normandy and MacArthur's mom's secret pie recipe will both be equally likely to be "decyphered". It will be impossible to guess at which one is the original message.

    Now let's say that Eve listens in on the sending of the original key to Bob. Every time she guesses which polarization of photon Alice sent Bob, she destroys the photon. If she guesses right, she can reproduce the photon and send the fake one back to Bob, and neither Alice nor Bob are the wiser. But if she guesses wrong she won't know the polarization of the photon, and she'll have to guess. There's a 50/50 chance she'll guess wrong, and a 50/50 chance Bob will guess right. Meaning all total there's a 1 in 8 chance every time Eve listens in that she'll cause an error, and half that (1 in 4) that the error will be reflected in the final key.

    So to combat this, Alice and Bob publicly share a random subset of their key. If they both used the same polarization, but got different bits, it means either someone was listening in and guessed wrong, or there's something faulty with the transmission. But for safety reasons they would always assume that someone was listening in.

    Assuming the bits they chose to share were chosen at random, there's a 1 - (3/4)^n chance that Eve will remain undetected. Where n is the number of bits in the key they decided to share. If they share upwards of a hundred bits, the chance of Eve being undetected drops to a number astronomically low.

    The probabilities get harder to calculate if Eve doesn't listen in on every single bit being transferred. But that also means she has to guess at parts of the key, and the final decrypted message will contain errors. So there's probably some sweet spot for the number of bits to compare based on some given message length.

    If Alice and Bob detect Eve, Alice doesn't transfer the encrypted message. They can try again or give up. But Eve can never know the key and trick Alice into thinking she doesn't.
    Reply With Quote  
     

  17. #16  
    Suspended
    Join Date
    Apr 2008
    Posts
    2,178
    Quote Originally Posted by Jarek Duda
    Quantum cryptography is made that You cannot just intercept the signal - in standard approach You don't even know what to measure.
    But ... one can install herself in the middle - tell Alice that she is Bob, Bob that she is Alice ... and finally get the transmitted messages.

    It's only another protocol - we cannot rely on it.
    Especially that we should be able to do that encrypted message is just safe.
    Ok - there is always human factor which greatly reduces safeness. But in good protection after firing a person, it should be enough to change codes.

    That is just ridiculous. If you keep complicating communications to a point that they cannot be deciphered. It will actually cause more death, destruction and eventual exposure of all involved, trying to hide something.

    It will cause poor communications, and a failed system, it will cause, a need for perpetual motion, to power the device. Might be worth while. Ha-ha.

    It will cause an unsafe unreliable system that fails and kills. No doubts about it. We have systems like that now. Very complicated, varying code systems. They can crack them in minutes.

    The reason why people like this stuff is that they know it is just silly. So even if it fails who cares, it is not important to real American citizens.

    A real Marine would tell the enemy, when he is coming, his numbers, and the equipment to expect. The enemy would either rethink his plan, commit suicide, or be slaughtered.

    All the counterintelligence, gives idiot leaders, the idea that we must be idiots too. And they might stand a chance against idiots like themselves.


    Sincerely,


    William McCormick
    Reply With Quote  
     

  18. #17  
    Suspended
    Join Date
    Apr 2008
    Posts
    2,178
    Quote Originally Posted by Jarek Duda
    Quantum cryptography is made that You cannot just intercept the signal - in standard approach You don't even know what to measure.
    But ... one can install herself in the middle - tell Alice that she is Bob, Bob that she is Alice ... and finally get the transmitted messages.

    It's only another protocol - we cannot rely on it.
    Especially that we should be able to do that encrypted message is just safe.
    Ok - there is always human factor which greatly reduces safeness. But in good protection after firing a person, it should be enough to change codes.

    A big complicated system, has so many back doors that there is no way to just take out one password and everything is well. Often a person involved in programing knows how to disable the entire network security for update or over haul, or backup and restore.

    I used to talk to accountants. And they are so believing in computer security. That I had to show them what a programer can do to information.
    I could press a button and reverse all the invoicing for any duration I wished, reprint the invoices, and the accounts receivable journal in a few minutes.
    And leave no trail.

    It only took a couple hours to write. Really cool to see work.


    There are often a lot of remote access passwords laying around. For when programers or workers forget theirs. With a phone call often you can get a name and password. A sexy voice, from a well sought after babe, to a hardened geek can often accomplish this.

    Some programs have security clearance. So if this fellow can run a certain program, he can get all the information he wants. Or change all the information he wants. And there is no record of it. Because it is done from a level, below the database. Some programs can be run from certain terminals anytime. Or are running a lot of the time. You can just halt the program and send it on a new mission.

    Or just call up some person, I would go with a woman in the company, imitate another man with a lot of clearance, home sick. And claim that the womans password came up in an incident, or log. It drew a flag. Probably something simple, failure to log out may be the cause. It could even be a technical problem. If I have to call, for your password, to confirm it is you, from the Network supervisor he makes note of it.

    I would of course, not want this to go into her annual rating, or pull it down. So I am in my weakened condition tracking down the password. That may be tied to another name, that would show it was not you, but a technical problem. That is what I am checking out. If it all gets cleared up, quietly no one gets a lowered review.

    I am sure I could get her password, and make some alterations rather quickly. Even add in another name and password. Even if they find it by the end of the day, I have the data base.

    I have friends and family in banking. I have no intention of doing this. And again, if you are not going to light a fire to the building, and take out power and gas to the state to cause chaos. Forget it.

    The Secret Service, United States Treasury department will be at your door with the local police, and perhaps, FBI, and or ATF (Alcohol,Tobacco, Firearms), Joint Federal Bureau of Investigation, and U.S. Treasury department team.



    Sincerely,


    William McCormick
    Reply With Quote  
     

  19. #18  
    Forum Junior
    Join Date
    Jul 2008
    Posts
    287
    Alice and Bob chooses randomly between two bases: '+' and 'x'. The measure gives some information iff they've chosen the same for given bit - later they use some classical channel to check which bits were read correctly and which should be sent again.

    So assume that Eve can intercept classical communication and cut the fiber wire and installed own devices on both sides - she can choose randomly own bases, compare with bases used by Alice and ask for resend ... and then use this protocol to send it to Bob.
    Reply With Quote  
     

  20. #19  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    Quote Originally Posted by Jarek Duda
    Alice and Bob chooses randomly between two bases: '+' and 'x'. The measure gives some information iff they've chosen the same for given bit - later they use some classical channel to check which bits were read correctly and which should be sent again.

    So assume that Eve can intercept classical communication and cut the fiber wire and installed own devices on both sides - she can choose randomly own bases, compare with bases used by Alice and ask for resend ... and then use this protocol to send it to Bob.
    Okay, I think I see what you mean. Essentially you're saying that if Bob and Alice want to communicate, they have no way of knowing if the person they're communicating with is really the one they want. Like if they're using a telephone, Eve can intercept the call and just parrot back to Bob what Alice says in real time.

    Yeah, that's still a limitation. There isn't any magic way of knowing who you're talking to. Just that whoever it is, you're talking to them securely. That is, there can only be a single listener for any given talker.
    Reply With Quote  
     

  21. #20  
    Forum Freshman Mach1ne's Avatar
    Join Date
    Feb 2009
    Location
    Miami, Florida
    Posts
    6
    Thanks for that synopsis numsgil, I actually understood it

    Lets say there was a way Alice could actually tell that she was communicating with Bob and not someone intercepting the communication. Would this then make it nearly impossible for Eve?
    Reply With Quote  
     

  22. #21  
    Forum Junior
    Join Date
    Jul 2008
    Posts
    287
    There can be for example attacks based on switching random number generator with some known deterministic one...

    Anyway authentication would require sending the key through the classical channel (can be intercepted) or using public key cryptography - in fact safeness of quantum cryptography relies on safeness of classical channel and classical cryptography... for me it's only a new way to get huge amount of money from rich paranoiac people...

    We should rather concentrate on creating just really safe classical cryptography instead.
    Reply With Quote  
     

  23. #22  
    Forum Freshman Mach1ne's Avatar
    Join Date
    Feb 2009
    Location
    Miami, Florida
    Posts
    6
    Hmm I see your point. It does seem to rely on the classic connections, but I feel like there must be some way to make the authentication process not rely on such a channel so it could not be intercepted...I just don't know what exactly
    Reply With Quote  
     

  24. #23  
    Forum Freshman holysword's Avatar
    Join Date
    Apr 2008
    Posts
    41
    Too long topic, I didn't had enought time to read it all yet - but I'll try to do soon.

    But as far as I know, quantum cryptography is PHYSICALLY IMPOSSIBLE to be broken, just because its IMPOSSIBLE to clone quantum information - a well know quantum property.

    We are not talking about "technological limitations". If you can clone a quantum state then we can have (even more) strange physical results...

    So, even if somehow someone steal an information throught a quantum channel, its really means "STEAL". As the information cannot be cloned, the receiver will not receive the data he was supposed to receive, and as so, the "attack" would be evident, limitating a lot of possibilities.

    Recent studies determines that maybe it can be possible to make an "aproximation" of a quantum state, but its a very recent field. Anyway, the "unclonable" status is good for security - and bad for a lot of other things.

    Sorry if someone already told that.
    "Nolite arbitrari quia venerim mittere pacem in terram non veni pacem mittere sed gladium"
    Yeshua Ha Mashiach
    Reply With Quote  
     

  25. #24  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    Quote Originally Posted by Mach1ne
    Thanks for that synopsis numsgil, I actually understood it

    Lets say there was a way Alice could actually tell that she was communicating with Bob and not someone intercepting the communication. Would this then make it nearly impossible for Eve?
    Not necessarily. Imagine that Eve installs a keylogger on Alice's computer. As Alice types in her message, the keylogger will record the keystrokes and Eve will have the message.

    Or maybe Eve installs a secret camera in Bob's house with a view of his computer screen. She'll be able to read the message as Bob does.

    So there's always ways to spy on people. This just makes it less convenient
    Reply With Quote  
     

  26. #25  
    Suspended
    Join Date
    Apr 2008
    Posts
    2,178
    Quote Originally Posted by Mach1ne
    Hmm I see your point. It does seem to rely on the classic connections, but I feel like there must be some way to make the authentication process not rely on such a channel so it could not be intercepted...I just don't know what exactly
    Verizon uses a timed security code, that expires every fifteen seconds. Generated by a device running in sink with the base security system. I installed one of the first ONT (Optical Network Terminal) systems in my area, with the two fellows, from the phone company, at my house. It was really interesting. They were new at installing them. So we all got an education.

    In order to initiate the network, the code input must match the security devices code. That matches the home base code every fifteen seconds.

    The truth is that the optical system, is pretty cool because there is a certain amount of location involved with the system. Meaning the line has a physical location. It goes from your house to the phone company.

    I believe Verizon truly treats your personal privacy and security, with the deepest respect. They seem to be more concerned for the system, rather then just the money. They seem to be long term oriented. And not just going with fads to make money. However today with the quickly changing hardware schemes it is hard to set hard unchanging standards.

    I have probably sold a million Verizon systems, after seeing the care they took to install the system. Very unlike the Cable Vision system. Cable Vision basically said, we just rent the cable, our software problem, is your software problem.

    Verizon said "Lets make this work". And spent many hours making it just right. I have never lost Internet service in almost two years. I did have a modem power supply go bad, it was sent to me the next day. They said they noted they had a problem with the power supply, and fixed it, in later models.

    But no matter what you do, exacting duplication of information, and communication. Requires exacting, recordable, repeatable, information to be sent over some communication system. Nothing is totally secure.



    Sincerely,


    William McCormick
    Reply With Quote  
     

  27. #26  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    Quote Originally Posted by William McCormick
    Nothing is totally secure...
    ... except quantum cryptography.
    Reply With Quote  
     

  28. #27  
    Forum Freshman thedrunk's Avatar
    Join Date
    May 2009
    Posts
    48
    SKY NET !

    I am enjoying this thread for I jsut recently learned abotu the Quantom Computer and am trying to find all information I can reguarding it and its possibilities.

    If any one would talk to me more about the whole QC subject please I am trying to write up a fictional explanation using QC as a basis for sci-fi computers and robotics.

    but if the QC can do E16 billion computations a second all it needs is a data base with all languages in order to compleate a brute for attack on any thing. how many different words are there in the world? live languages dead lang's trying billions*trillions of them per micro second all at once is bound to break one thing if not many.


    Think about this scenario. a QC the size of a Qubit vs a bit. take the size differance and make a fictional computer the size of todays, say a quard core QC with max ram and a data base of all known written lang. if the computer can at a 128bit do those things think of a 1024 trillion Qbit system? how many computations can it do per word, per lang, at x amount of encryption? the possibilites of a QC are endless. what scares me is after a QC it will be a string computer....................
    this might sound unprofessional well it is. it's just the thought of comparing the Qbit to BIT, Qbit's form what i under stand can not only do a 1 and a 0 but a combination of 1 and 0 at the same times its doing both a 1 and 0. where a regualr computer can only do a 1 or 0 never both at same time. think if im wrong please correct me I have only had some google knowledge about QC...
    Reply With Quote  
     

  29. #28  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    Quote Originally Posted by thedrunk
    but if the QC can do E16 billion computations a second all it needs is a data base with all languages in order to compleate a brute for attack on any thing. how many different words are there in the world? live languages dead lang's trying billions*trillions of them per micro second all at once is bound to break one thing if not many.
    First, It's not that a QC is faster. It's just... different. Some problems that are hard for a classic computer are easy for a QC. Easy in the sense of lower Big O notation, if you're familiar with that. Factoring integers famously. I also think, but could be wrong, that some problems which are easy for classical computers might be hard for a QC, but I can't find hard data on that.

    Also, QC can't do anything that a classic computer can't (such as the halting problem).

    I think it's probable that future computers will be a hybrid of classical computers and quantum computers. Even if quantum computers are entirely better all around, the technology for classic computer hardware is extremely advanced.

    Second, breaking codes isn't always a matter of finding the right password, and when it is, it's almost never a case of finding an actual "word". If you read back in this thread I explain how quantum cryptography works. In a world where quantum computers are common, that would be the cryptographic method of choice. You can never "break" a quantum encrypted message, because all possible messages are equally likely to be "decrypted" from the message. Not even with a QC. It's impossible.

    However old (ie: current) encrypted messages would be pretty easy, because most current encryption methods rely on integers being hard to factor. And if you have a QC it's relatively easy (in that it's possible to do in a "reasonable" period of time instead of years).

    So from a fiction point of view, a QC means you can break encryptions from, say, an old 20th or early 21st century abandoned military base. Or if you're an alien with a QC, you can break encryptions on present day Earth. But it doesn't mean that in the future encryption is impossible. It just changes its form. And of course there are always human error type ways to get at information you shouldn't have.
    "A witty saying proves nothing." - Voltaire
    Reply With Quote  
     

  30. #29  
    Forum Freshman thedrunk's Avatar
    Join Date
    May 2009
    Posts
    48
    Quote Originally Posted by Numsgil
    Quote Originally Posted by thedrunk
    but if the QC can do E16 billion computations a second all it needs is a data base with all languages in order to compleate a brute for attack on any thing. how many different words are there in the world? live languages dead lang's trying billions*trillions of them per micro second all at once is bound to break one thing if not many.
    First, It's not that a QC is faster. It's just... different. Some problems that are hard for a classic computer are easy for a QC. Easy in the sense of lower Big O notation, if you're familiar with that. Factoring integers famously. I also think, but could be wrong, that some problems which are easy for classical computers might be hard for a QC, but I can't find hard data on that.

    Also, QC can't do anything that a classic computer can't (such as the halting problem).

    I think it's probable that future computers will be a hybrid of classical computers and quantum computers. Even if quantum computers are entirely better all around, the technology for classic computer hardware is extremely advanced.

    Second, breaking codes isn't always a matter of finding the right password, and when it is, it's almost never a case of finding an actual "word". If you read back in this thread I explain how quantum cryptography works. In a world where quantum computers are common, that would be the cryptographic method of choice. You can never "break" a quantum encrypted message, because all possible messages are equally likely to be "decrypted" from the message. Not even with a QC. It's impossible.

    However old (ie: current) encrypted messages would be pretty easy, because most current encryption methods rely on integers being hard to factor. And if you have a QC it's relatively easy (in that it's possible to do in a "reasonable" period of time instead of years).

    So from a fiction point of view, a QC means you can break encryptions from, say, an old 20th or early 21st century abandoned military base. Or if you're an alien with a QC, you can break encryptions on present day Earth. But it doesn't mean that in the future encryption is impossible. It just changes its form. And of course there are always human error type ways to get at information you shouldn't have.
    My question would be what about a tri-breed system? DNA, QC, and standard?


    What is the classification of a standard computer?
    Reply With Quote  
     

  31. #30  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    I don't know enough about DNA computers to know what their strengths/weaknesses are. But yeah, why not? :P

    What do you mean by "classification"?
    "A witty saying proves nothing." - Voltaire
    Reply With Quote  
     

  32. #31  
    Forum Freshman thedrunk's Avatar
    Join Date
    May 2009
    Posts
    48
    Quote Originally Posted by Numsgil
    I don't know enough about DNA computers to know what their strengths/weaknesses are. But yeah, why not? :P

    What do you mean by "classification"?
    Classification as in anolog ditigital, or what. what is the standard term for a standard computer liek we have today?


    well from what i under stand about DNA computers is that they duplicate till they solve the problem the recycle the reproduced componets in order to save room.

    I am looking in to all this at once and its a CF of information to sort through and figuring out reliable sources and what not.
    Reply With Quote  
     

Bookmarks
Bookmarks
Posting Permissions
  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •