Polynomial secret sharing schemes generalize the linear ones by adding more expressivity and giving room for efficiency improvements. In these schemes, the secret is an element of a finite field, and the shares are obtained by evaluating polynomials on the secret and some random field elements, i.e., for every party there is a set of polynomials that computes the share of the party. Notably, for general access structures, the best known polynomial schemes have better share size than the best known linear ones. This work investigates the properties of the polynomials and the fields that allow these efficiency gains and aims to characterize the access structures that benefit from them.
We focus first on ideal schemes, which are optimal in the sense that the size of each share is the size of the secret. We prove that, under some restrictions, if the degrees of the sharing polynomials are not too big compared to the size of the field, then its access structure is a port of an algebraic matroid. This result establishes a new connection between ideal schemes and matroids, extending the known connection between ideal linear schemes and linearly representable matroids.
For general polynomial schemes, we extend these results and analyze their privacy and correctness. Additionally, we show that given a set of polynomials over a field of large characteristic, one can construct linear schemes that realize the access structure determined by these polynomials; as a consequence, polynomial secret sharing schemes over these fields are not stronger than linear schemes.
While there exist ports of algebraic matroids that do not admit ideal schemes, we demonstrate that they always admit schemes with statistical security and information ratio tending to one. This is achieved by considering schemes where the sharings are points in algebraic varieties.
