Encryption is dead

[Pu239 Stuchtiger]

A team of people from India (if I remember correctly) developed an algorithm that determines whether a number is prime. So much for encryption!

[nbk2000]

Christ! Get it straight.

It's not determining primality that would be the "end of encryption", but rather factoring. That is, determining which two prime numbers combined, make up the the product that is the encryption key.

THIS is the hard part.

Read "applied cryptography" 2nd ed. for a better understanding of what's involved. Or, you could read the PGP FAQ PDF that comes included with the program. It gives a good overview.

[Caacrinolaas]

The algorithm could help encryption insteed of kill it since all the encryption programs that use prime numbers must use an algorithm to determine if the number they are using for encryption is an prime, the algorithm PGP use is in fact not 100% secure in it's determination of prime numbers. If i remember my numbers right i calculated that somewhere around 2 people in the world use an number for encryption that isn't a prime... anyhow what will really be the end of the prime based encryption is quantum computers, those nice lil suckers have managed to factor 15 to its prime 3 and 5. :)

[PYRO500]

Hey, don't forget that it solved that problem in 0 time!

[Marvin]

Try also reading "The PGP attack FAQ", its by the same guy that last I saw was editing phrack, and is very good indeed. There are a lot of versions floating round the net so it shouldnt be hard to find.

Even QC's are only of threat mainly to public key cryptography, and its doubtful it will kill that completely. Assuming one can be built large enough to solve 512bit RSA for example, that would be of little or no use in solving larger numbers. Its possible the goverment would restrict the largest QC's that can be made for the public and attempt to keep one step ahead of generally known unsafe RSA sizes.

Private key cryptography can be engineered to be perfectly safe from a QC of any size by including redundancy into the alg. By encrypting based on the output of a one way hash rather than a symetric block cipher, there are multiple solutions for a key with any plaintext/ciphertext pair and this prevents the wavefunction collapsing to any one solution. Incorperate enough redundancy into the system, eg 128bits and you completly prevent any attempt to collapse the wavefunction by guessing the state of bits beforehand. Add to this a hardcoded, randomly produced second layer of encryption, where the code for that alg is embedded in the first layer of encryption, force the decryption check to be small, say 64bits, and that prevents a plaintext attack on the known alg. Note that the second layer is completely redundant in the case of the primary encryption system being broken, it has the code and the key to simply unravel itself and is present only to hinder a multiple plaintext attack being done at the same time on a QC that would need to be massive anyway.

In the event of any of this, the key is only as safe as the easiest way someone else can find it, Bugs in the keyboard or computer, passive snooping methods or viri/trojens will always be easier, and dont require an annoucement that the goverment has been able to see your secrets all along. Any encrpytion is better than none, and most people are much more likley to be gotcha'd by failing to encrypt the sensitive information, than have a system broken speicifically for the information on their HD.

Appropriately enough, the quantum mechanics in a chemistry degree is an excelent foundation for understanding quantum computing.

[Anarchist]

just to let you know pu239, not all encryption is based off of prime numbers. There are many other algorithims that have nothing to do with primes. As nbk suggested get applied cryptopgraphy it is the single best book on computer(and non-computer) related cryptography in print today.