Secure two-party protocols are of significant research and application value. All-or-Nothing Disclosure of Secrets (ANDOS) is such a kind of cryptographic task. It involves two parties, a vendor and a buyer. The vendor, say Alice, who disposes of several secrets and is willing to sell any of them to the buyer, say Bob, with the guarantee that no information about the other secrets will be obtained. Furthermore, Bob can freely choose his secret and wants to ensure that Alice can obtain on information about which of her secrets he has picked. In this paper, we present a new quantum ANDOS scheme which achieves the same functionality, but which is of unconditional security.
A vendor has several secrets which he regards as of equal worth and importance. He is willing to sell any of these secrets to a buyer. The bargain is said to be secure when the following two conditions are satisfied: 1) the buyer can only obtain one secret if he pays for only one secret, and 2) the vendor will not be able to find out which secret the buyer picked. This cryptographic problem is customarily called all-or-nothing disclosure of secrets (ANDOS). Previous ANDOS schemes have a few defects: neither the number of the secrets is restricted within narrow limits, nor the security is desirable. In this paper, we propose two new and novel ANDOS protocols in quantum computing environment. The first protocol is shown to be as secure as a classical one. The second protocol is based on unambiguous state discrimination, which is proven to be secure in quantum world. Moreover, all our schemes have no limitation on the number of the secrets to be sold.
Alice disposes of some number of secrets. She is willing to disclose one of them to Bob. Although she agrees to let him choose which secret he wants, she is not willing to allow him to gain any information on more than one secret. On the other hand, Bob does not want Alice to know which secret he wishes. An all-or-nothing disclosure is one by which, as soon as Bob has gained any information whatsoever on one of Alice's secrets, he has wasted his chances to learn anything about the other secrets. We assume that Alice is honest when she claims to be willing to disclose one secret to Bob (i.e. she is not about to send junk). The only cheating Alice is susceptible of trying is to figure out which secret is of interest to Bob. We address the following question from an information theoretic point of view: what is the most elementary disclosure problem? The main result is that the general all-or-nothing disclosure of secrets is equivalent to a much simpler problem, which we call the two-bit problem.