Thesis (Ph. D.)--Auburn University, 2009. / Abstract. Includes bibliographical references (p. 51-52).

Parmi les différents critères qu'une fonction booléenne doit satisfaire en cryptographie, on s'intéresse à la non-linéarité. Pour une fonction booléenne donnée, cette notion mesure la distance de Hamming qui la sépare des fonctions de degré au plus 1. C'est un critère naturel pour évaluer la complexité d'une fonction cryptographique, celle-ci ne devant pas admettreune approximation qui soit simple, comme par une fonction de degré 1, ou plus généralement une fonction de bas degré. Ainsi, il est important de considérer plus généralement, la non-linéarité d'ordre supérieur, qui pour un ordre donné r, mesure la distance d'une fonction donnée à l'ensemble des fonctions de degré au plus r. Cette notion est également importante pour les fonctions vectorielles, i.e., celles à plusieurs sorties. Quand le nombre de variables est grand, presque toutes les fonctions ont une non-linéarité (d'ordre 1) voisine d'une certaine valeur, assez élevée. Dans un premier travail, on étend ce résultat à l'ordre 2. Cette méthode qui consiste à observer comment les boules de Hamming recouvrent l'hypercube des fonctions booléennes, nous conduit naturellement vers une borne de décodage théorique des codes de Reed-Muller d'ordre 1, coïncidant au même endroit où se concentre la non-linéarité de presque toutes les fonctions ; une approche nouvelle pour un résultat pas entièrement nouveau. On étudie aussi la non-linéarité des fonctions vectorielles. On montre avec une approche différente, que le comportement asymptotique est le même que celui des fonctions booléennes: une concentration de la non-linéarité autour d'une valeur assez élevée. / Among the different criteria that a Boolean function must satisfy in symmetric cryptography, we focus on the nonlinearity of these. This notion measures the Hamming distance between a given function and the set of functions with degree at most 1. It is a natural criterion to evaluate the complexity of a cryptographic function that must not have a simple approximation as by a function of degree 1, or more generally, a function of low degree. Hence, it is important to consider the higher order nonlinearity, which for a given order r, measures the distance between a given function and the set of all functions of degree at most r. This notion is equally important for multi-output Boolean functions. When the number of variables is large enough, almost all Boolean functions have nonlinearities lying in a small neighbourhood of a certain high value. We prove that this fact holds when considering the second-order nonlinearity. Our method which consists in observing how the Hamming balls pack the hypercube of Boolean functions led quite naturally to a theoretical decoding bound for the first-order Reed-Muller code, coinciding with the concentration point of the nonlinearity of almost all functions. This was a new approach for a result which is not entirely new. We also studied the nonlinearity of multi-output functions. We proved with a different approach, that the asymptotic behaviour of multi-output functions is the same as the single-output ones: a concentration of the nonlinearity around a certain large value.

The importance of quantum error correction in paving the way to build a practical quantum computer is no longer in doubt. Despite the large body of literature in quantum coding theory, many important questions, especially those centering on the issue of "good codes" are unresolved. In this dissertation the dominant underlying theme is that of constructing good quantum codes. It approaches this problem from three rather different but not exclusive strategies. Broadly, its contribution to the theory of quantum error correction is threefold. Firstly, it extends the framework of an important class of quantum codes - nonbinary stabilizer codes. It clarifies the connections of stabilizer codes to classical codes over quadratic extension fields, provides many new constructions of quantum codes, and develops further the theory of optimal quantum codes and punctured quantum codes. In particular it provides many explicit constructions of stabilizer codes, most notably it simplifies the criteria by which quantum BCH codes can be constructed from classical codes. Secondly, it contributes to the theory of operator quantum error correcting codes also called as subsystem codes. These codes are expected to have efficient error recovery schemes than stabilizer codes. Prior to our work however, systematic methods to construct these codes were few and it was not clear how to fairly compare them with other classes of quantum codes. This dissertation develops a framework for study and analysis of subsystem codes using character theoretic methods. In particular, this work established a close link between subsystem codes and classical codes and it became clear that the subsystem codes can be constructed from arbitrary classical codes. Thirdly, it seeks to exploit the knowledge of noise to design efficient quantum codes and considers more realistic channels than the commonly studied depolarizing channel. It gives systematic constructions of asymmetric quantum stabilizer codes that exploit the asymmetry of errors in certain quantum channels. This approach is based on a Calderbank- Shor-Steane construction that combines BCH and finite geometry LDPC codes.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / In this work we present the determination of the second Hamming weight of generalized Reed- Muller codes in most cases (see Teorema 4.6). Our main reference is [13], although we have also used results from [3] and [5]. In the first chapter we describe finite fields e we show how they can be constructed. In chapter 2 we present the basics of coding theory. We define what are error correcting codes, the Hamming metric, the parameters of a code, the equivalence of codes through the concept of isometry, and we briefly present generalized Reed-Muller codes and their parameters. In chapter 3 we present some results from Grobner bases theory and the definition of Affine Cartesian codes, which generalize the generalized Reed-Muller codes. we use tools from Grobner bases theory to determine the dimension and the minimum distance of Affine Cartesian codes. We finish our work in chapter 4, with the determination of the second Hamming weight for generalized Reed-Muller codes in most cases. / Nesse trabalho apresentamos o cálculo do segundo peso de Hamming de códigos de Reed-Muller generalizados na maioria dos casos (v. Teorema 4.6). Nossa referência principal sera [13], embora tenhamos utilizado também resultados de [3] e [5]. No primeiro capítulo descrevemos os corpos finitos e mostramos como podem ser construídos. No capítulo 2 apresentamos os conceitos básicos da teoria de códigos. Nele, definimos o que são os códigos corretores de erros, a métrica de Hamming, os parâmetros de um código, a equivalência de códigos através da noção de isometria, bem como uma breve apresentação dos códigos de Reed-Muller generalizados e seus parâmetros. No capítulo 3 sao apresentados alguns resultados da teoria de Bases de Grobner e a definição dos Códigos Cartesianos Afins, que são uma generalização dos códigos de Reed-Muller generalizados. Usamos ferramentas da teoria de bases de Grobner para determinar a dimensão e distância mínima de Códigos Cartesianos Afins. Para finalizar nosso trabalho, no capítulo 4 determinamos o segundo peso de Hamming do Código de Reed-Muller generalizado na maioria dos casos. / Mestre em Matemática

Contrairement aux protocoles cryptographiques fondés sur la théorie des nombres, les systèmes de chiffrement basés sur les codes correcteurs d’erreurs semblent résister à l’émergence des ordinateurs quantiques. Un autre avantage de ces systèmes est que le chiffrement et le déchiffrement sont très rapides, environ cinq fois plus rapide pour le chiffrement, et 10 à 100 fois plus rapide pour le déchiffrement par rapport à RSA. De nos jours, l’intérêt de la communauté scientifique pour la cryptographie basée sur les codes est fortement motivé par la dernière annonce de la “National Institute of Standards and Technology" (NIST), qui a récemment initié le projet intitulé “Post-Quantum cryptography Project". Ce projet vise à définir de nouveaux standards pour les cryptosystèmes résistants aux attaques quantiques et la date limite pour la soumission des cryptosystèmes à clé publique est fixée pour novembre 2017. Une telle annonce motive certainement à proposer de nouveaux protocoles cryptographiques basés sur les codes, mais aussi à étudier profondément la sécurité des protocoles existants afin d’écarter toute surprise en matière de sécurité. Cette thèse suit cet ordre d’idée en étudiant la sécurité de plusieurs protocoles cryptographiques fondés sur la théorie des codes correcteurs d’erreurs. Nous avons commencé par l’étude de la sécurité d’une version modifiée du cryptosystème de Sidelnikov, proposée par Gueye et Mboup [GM13] et basée sur les codes de Reed-Muller. Cette modification consiste à insérer des colonnes aléatoires dans la matrice génératrice (ou de parité) secrète. La cryptanalyse repose sur le calcul de carrés du code public. La nature particulière des codes de Reed-Muller qui sont définis au moyen de polynômes multivariés binaires, permet de prédire les valeurs des dimensions des codes carrés calculés, puis permet de récupérer complètement en temps polynomial les positions secrètes des colonnes aléatoires. Notre travail montre que l’insertion de colonnes aléatoires dans le schéma de Sidelnikov n’apporte aucune amélioration en matière de sécurité. Le résultat suivant est une cryptanalyse améliorée de plusieurs variantes du cryptosystème GPT qui est un schéma de chiffrement en métrique rang utilisant les codes de Gabidulin. Nous montrons qu’en utilisant le Frobenius de façon appropriée sur le code public, il est possible d’en extraire un code de Gabidulin ayant la même dimension que le code de Gabidulin secret mais avec une longueur inférieure. Le code obtenu corrige ainsi moins d’erreurs que le code secret, mais sa capacité de correction d’erreurs dépasse le nombre d’erreurs ajoutées par l’expéditeur et par conséquent, un attaquant est capable de déchiffrer tout texte chiffré, à l’aide de ce code de Gabidulin dégradé. Nos résultats montrent qu’en fin de compte, toutes les techniques existantes visant à cacher la structure algébrique des codes de Gabidulin ont échoué. Enfin, nous avons étudié la sécurité du système de chiffrement de Faure-Loidreau [FL05] qui est également basé sur les codes de Gabidulin. Inspiré par les travaux précédents et, bien que la structure de ce schéma diffère considérablement du cadre classique du cryptosystème GPT, nous avons pu montrer que ce schéma est également vulnérable à une attaque polynomiale qui récupère la clé privée en appliquant l’attaque d’Overbeck sur un code public approprié. Comme exemple, nous arrivons en quelques secondes à casser les paramètres qui ont été proposés comme ayant un niveau de sécurité de 80 bits. / Contrary to the cryptosystems based on number theory, the security of cryptosystems based on error correcting codes appears to be resistant to the emergence of quantum computers. Another advantage of these systems is that the encryption and decryption are very fast, about five times faster for encryption, and 10 to 100 times faster for decryption compared to RSA cryptosystem. Nowadays, the interest of scientific community in code-based cryptography is highly motivated by the latest announcement of the National Institute of Standards and Technology (NIST). They initiated the Post-Quantum cryptography Project which aims to define new standards for quantum resistant cryptography and fixed the deadline for public key cryptographic algorithm submissions for November 2017. This announcement motivates to study the security of existing schemes in order to find out whether they are secure. This thesis thus presents several attacks which dismantle several code-based encryption schemes. We started by a cryptanalysis of a modified version of the Sidelnikov cryptosystem proposed by Gueye and Mboup [GM13] which is based on Reed-Muller codes. This modified scheme consists in inserting random columns in the secret generating matrix or parity check matrix. The cryptanalysis relies on the computation of the square of the public code. The particular nature of Reed-Muller which are defined by means of multivariate binary polynomials, permits to predict the values of the dimensions of the square codes and then to fully recover in polynomial time the secret positions of the random columns. Our work shows that the insertion of random columns in the Sidelnikov scheme does not bring any security improvement. The second result is an improved cryptanalysis of several variants of the GPT cryptosystem which is a rank-metric scheme based on Gabidulin codes. We prove that any variant of the GPT cryptosystem which uses a right column scrambler over the extension field as advocated by the works of Gabidulin et al. [Gab08, GRH09, RGH11] with the goal to resist Overbeck’s structural attack [Ove08], are actually still vulnerable to that attack. We show that by applying the Frobeniusoperator appropriately on the public key, it is possible to build a Gabidulin code having the same dimension as the original secret Gabidulin code, but with a lower length. In particular, the code obtained by this way corrects less errors than thesecret one but its error correction capabilities are beyond the number of errors added by a sender, and consequently an attacker is able to decrypt any ciphertext with this degraded Gabidulin code. We also considered the case where an isometrictransformation is applied in conjunction with a right column scrambler which has its entries in the extension field. We proved that this protection is useless both in terms of performance and security. Consequently, our results show that all the existingtechniques aiming to hide the inherent algebraic structure of Gabidulin codes have failed. To finish, we studied the security of the Faure-Loidreau encryption scheme [FL05] which is also a rank-metric scheme based on Gabidulin codes. Inspired by our precedent work and, although the structure of the scheme differs considerably from the classical setting of the GPT cryptosystem, we show that for a range of parameters, this scheme is also vulnerable to a polynomial-time attack that recovers the private key by applying Overbeck’s attack on an appropriate public code. As an example we break in a few seconds parameters with 80-bit security claim.

Employing an error control code is one of the techniques to reduce the Peak-to-Average Power Ratio (PAPR) in an Orthogonal Frequency Division Multiplexing system; a well known class of such codes being the cosets of Reed-Muller codes. In this thesis, classes of such coset-codes of arbitrary linear codes are considered. It has been proved that the size of such a code can be doubled with marginal/no increase in the PAPR. Conditions for employing this method iteratively have been enunciated. In fact this method has enabled to get the optimal coset-codes. The PAPR of the coset-codes of the extended codes is obtained from the PAPR of the corresponding coset-codes of the parent code. Utility of a special type of lengthening is established in PAPR studies