The aim of this research is the efficient cryptanalysis of the Shrinking Generator through its characterization by means of Linear Hybrid Cellular Automata. This paper describes a new known-plaintext attack based on the computation of the characteristic polynomials of sub-automata and on the generation of the Galois field associated to one of the Linear Feedback Shift Registers components of the generator. The proposed algorithm allows predicting with absolute certainty, many unseen bits of the keystream sequence, thanks to the knowledge of both registers lengths, the characteristic polynomial of one of the registers, and the interception of a variable number of keystream bits.

An efficient computational Zero-Knowledge Proof of Knowledge whose security relies on the NP-completeness of the Independent Set Problem is presented here. The proposed algorithm is constructed from a bit commitment scheme based on the hardness of the Discrete Logarithm Problem, which guarantees the fulfillment of soundness, completeness and computational zero-knowledge properties, and allows avoiding the use of the Graph Isomorphism Problem, which is present in every known Zero-Knowledge Proofs for the Independent Set Problem.

The aim of this work is to investigate the possibility of designing zero-knowledge identification schemes based on hard-on-average problems. It includes a new two-party identification protocol whose security relies on a discrete mathematics problem classified as DistNP-Complete under the average-case analysis, the so-called Distributional Matrix Representability Problem. Thanks to the use of the search version of the mentioned problem, the zero-knowledge property is formally proved by black-box simulation, and consequently the security of the proposed scheme is actually guaranteed. Furthermore, with the proposal of a new zero-knowledge proof based on a problem never used before for this purpose, the set of tools for designing cryptographic applications is enlarged.

This work addresses the critical problem of authentication in mobile ad hoc networks. It includes a new approach based on the Zero-Knowledge cryptographic paradigm where two different security levels are defined. The first level is characterized by the use of an NP-complete graph problem to describe an Access Control Protocol, while the highest level corresponds to a Group Authentication Protocol based on a hard-on-average graph problem. The main goal of the proposal is to balance security strength and network performance. Therefore, both protocols are scalable and decentralized, and their requirements of communication, storage and computation are limited.