Quantum-noise randomized ciphers

Abstract

We review the notion of a classical random cipher and its advantages. We sharpen the usual description of random ciphers to a particular mathematical characterization suggested by the salient feature responsible for their increased security. We describe a concrete system known as ␣ and show that it is equivalent to a random cipher in which the required randomization is affected by coherent-state quantum noise. We describe the currently known security features of ␣ and similar systems, including lower bounds on the unicity distances against ciphertext-only and known-plaintext attacks. We show how ␣ used in conjunction with any standard stream cipher such as the Advanced Encryption Standard provides an additional, qualitatively different layer of security from physical encryption against known-plaintext attacks on the key. We refute some claims in the literature that ␣ is equivalent to a nonrandom stream cipher.

On the security of Y-00 under fast correlation and other attacks on the key

“…A fast correlation attack (FCA) was presented that was shown to succeed by simulations for moderate signal levels when the ENC box in Y-00 is a LFSR (linear feedback shift register) of a few taps and length up to 32. Even though such Y-00 is already insecure against what we call assisted brute-force search [9] due to the small seed key size |K| ≤ 32, such FCA is of interest as it brings forth the whole issue of Y-00 seed key security against similar and other attacks.…”

Exposed-key weakness of αη

“…In this paper, we show that αη has an additional disadvantage that current classical ciphers do not have: transmission of the "encrypted" states actually leaks information about the key to an eavesdropper, even if that eavesdropper has no information about the message. Such a weakness has been independently described briefly in [6], and in more depth in [7] and [8]; here, we calculate information loss and estimate a bound on the efficacy of explicit attacks such as [7]. In the remainder of this paper, we will describe the αη scheme, show that in practice there is no advantage created for Bob over Eve via quantum limits on measurement, and estimate how much information Eve can learn about the key from Alice's transmission.…”

“…Another advantage of αη stated in [6] is that it gains some security due to the physical nature of the states being sent: it may in practice be difficult to perform the measurements needed to eavesdrop on a channel with this encoding. However, since an effective eavesdropping strategy is to employ a heterodyne or dual-homodyne measurement, we do not see a great difference in the practical difficulty of eavesdropping from the difficulty of the legitimate receiver, or from the receiver of any coherent communication system.…”

“…In the remainder of this paper, we will describe the αη scheme, show that in practice there is no advantage created for Bob over Eve via quantum limits on measurement, and estimate how much information Eve can learn about the key from Alice's transmission. We will also discuss a technique given in [9] and [6], Deliberate Signal Randomization, and how our analysis applies there.…”

“…In [6,8], it is argued that a technique that is called Deliberate Signal Randomization (DSR) will serve to add information-theoretic security to the key. This technique simply involves Alice sending a random state in the half-plane around the state she would send under the non-DSR version given above.…”