Unconditional security against memory-bounded adversaries.

From MaRDI portal





To prove security of a cryptographic system two approaches are commonly used. In the first one, the security of the cryptosystem is based on the difficulty of an underlying computational problem. An alternative approach is based on the notion of information-theoretic (or unconditional) security under the assumption that an adversary's computational power is not limited. The traditional information-theoretic definition of perfect security is well known to lead to perfectly secure but impractical cryptosystems. However, it has been shown recently that by relaxing or removing some assumptions of the original model, a practical provably secure cryptosystem becomes possible. So far two such modifications were proposed, namely relaxing the requirement that perfect security means complete independence between the plaintext and the cryptanalyst's knowledge, or removing the assumption that the adversary receives exactly the same information as the legitimate users. In the paper yet another modification is introduced and shown to allow the realization of unconditionally secure encryption. The key assumption of the new model is that an adversary's memory capacity is limited while nothing is assumed about her computing power. Particularly, it is shown that assuming that all parties can receive a random bit string of length larger than the adversary's memory capacity, practical unconditionally secure private-key encryption and key-agreement (by public discussion) protocol can be implemented. It must be stressed, however, that the proposed scheme requires that a random bit string must be generated and broadcast at a very high rate. An additional distinctive feature of the new approach is that a future breaking of the basic assumption (i.e. future increase of an adversary's storage capacity) does not yield to the breaking of messages encrypted earlier. Some other features as well as possible drawbacks of the approach are also discussed. For the entire collection see [Zbl 0870.00047].




Cited in
(44)








This page was built for publication: Unconditional security against memory-bounded adversaries.

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4364559)