Black-Box Reductions and Separations in Cryptography

Cryptographic constructions of one primitive or protocol from another one usually come with a reductionist security proof, in the sense that the reduction turns any adversary breaking the derived scheme into a successful adversary against the underlying scheme. Very often the reduction is black-box in the sense that it only looks at the input/output behavior of the adversary and of the underlying primitive. Here we survey the power and the limitations of such black-box reductions, and take a closer look at the recent method of meta-reductions.