koning@dec-koning.UUCP (04/22/86)
>The simplest encryption which is secure is the one time pad, right? >The problem is that the volume of key is as big as the volume of >message, right? Would the availability of enormous volumes of digital >information permit convenient construction (at a distance) of a new >effective one time pad from the separate (secure) transmission of a >few thousand bits to be used in selecting larger supplies of bits >from one or several CD ROMs or simply compact disks of music? Since >I am not aware of them, I would like to hear about techniques for >calculating the difficuly of decrypting for any non-RSA scheme. secure cryptosystem? ONLY (theoretically, rather than computationally) secure cryptosystem? Anyway, the security of one-time pad depends on the requirement that the pad be random. Music is not random at all and in particular adjacent samples will be highly correlated since the sample rate is 44 kHz and the amount of energy at the high end of the audio spectrum is rather low. (Tweeters depend on that...) On the other hand, compact disks WOULD make a marvelous way to store true one-time pads, except for the problem that the mastering cost is high and you obviously will make only one or at most two copies of each master... which is a problem that will go away with high-density write-once disks. Come to think of it, you could write the "write-once" disks a second time to obliterate the part of the one-time pad you've just used, and secure what you've just sent against the risk of your key disk being captured. paul
koning@dec-koning.UUCP (04/22/86)
>The simplest encryption which is secure is the one time pad, right? >The problem is that the volume of key is as big as the volume of >message, right? Would the availability of enormous volumes of digital >information permit convenient construction (at a distance) of a new >effective one time pad from the separate (secure) transmission of a >few thousand bits to be used in selecting larger supplies of bits >from one or several CD ROMs or simply compact disks of music? Since >I am not aware of them, I would like to hear about techniques for >calculating the difficuly of decrypting for any non-RSA scheme. Isn't one-time pad the ONLY (theoretically, rather than computationally) secure cryptosystem? Anyway, the security of one-time pad depends on the requirement that the pad be random. Music is not random at all and in particular adjacent samples will be highly correlated since the sample rate is 44 kHz and the amount of energy at the high end of the audio spectrum is rather low. (Tweeters depend on that...) On the other hand, compact disks WOULD make a marvelous way to store true one-time pads, except for the problem that the mastering cost is high and you obviously will make only one or at most two copies of each master... which is a problem that will go away with high-density write-once disks. Come to think of it, you could write the "write-once" disks a second time to obliterate the part of the one-time pad you've just used, and secure what you've just sent against the risk of your key disk being captured. paul
tedrick@ernie.berkeley.edu (Tom Tedrick) (04/23/86)
>Isn't one-time pad the ONLY (theoretically, rather than computationally) >secure cryptosystem? "ONLY" is a tricky word. For example Giles Brassard has published some work on "Quantum Cryptography" where security depends on principles of quantum physics. One reference is "An Update on Quantum Cryptography", by Charles H. Bennet & Giles Brassard, pages 475-480, "Advances in Cryptology: Proceedings of Crypto 84", Springer Verlag Lecture Notes in Computer Science #196.