Covers and Logarithmic Signatures of Finite Groups in Cryptography
Pavol Svaba · DuEPublico (University of Duisburg-Essen) · 2011
After the first time Diffie and Hellmann [1] introduced the idea of separate keys, asymmetric cryptography has became increasingly developing. Many public key cryptosystems have been proposed, but only few of such systems remain unbroken. The most of them used nowadays are based on the perceived intractability of certain mathematical problems in very large, finite cyclic groups. In the late 1970's S. Magliveras started to investigate the use of special factorizations, called logarithmic signatures, of finite non-abelian groups in cryptography [2,3,4,5]. Later, Magliveras, Stinson and Tran van Trung [6] have done some preliminary work in creating two public key cryptosystems, MST1, based on logarithmic signatures, and MST2, based on another type of group coverings called [s,r]-meshes. Until now however, no practical realizations are known for MST1 or MST2. Recently, a new type of public key cryptosystem, called MST3 [7], has been developed on the basis of logarithmic signatures and random covers of finite non-abelian groups (i.e. factorization sequences in which blocks are constructed by sampling uniformly at random on the underlying group). For a possible realization of the generic version of this system, the Suzuki 2-groups have been suggested. %Due to their simple structure, these groups make it possible for studying the security of the scheme. The primary objective of this thesis is to show that cryptosystem MST3 can be realized with Suzuki 2-groups. To this question we can give an affirmative answer. There are several challenges in designing the practical realization of the scheme. The first problem is to efficiently generate covers for large groups and with good cryptographic properties. Showing the connection of this problem with the classical occupancy problem, we determine a bound for the probability that randomly chosen collection of group elements compose a cover. As a consequence, we solve the problem of generating random covers for arbitrary large groups. We also present several experimental computer results about covers and uniform covers for some alternating groups.Due to their simple structure, the Suzuki 2-groups enable us to study the security of the system and also provide an efficient implementation. In the first realization, a special class of canonical logarithmic signatures for elementary abelian 2-groups has been proposed as a basis for the key generation. These are easily constructed and allow highly efficient factorization. We provide an attack, showing that canonical signatures cannot be used to build secure realization of MST3 with Suzuki 2-groups. Motivated by the attack on the first realization, we propose a new variant with significant improvement, strengthening the system's security. For that purpose we re-design the set-up of the scheme and introduce a new class of fused transversal logarithmic signatures. These allow efficient factorization if we keep track of the transformations used to generate them. We present a thorough study of the security of the scheme by using heuristic and algebraic methods. We first determine the complexity for the lower bounds of conceivable direct attacks to recover the private key in terms of the size of the groups. These bounds give a hint of the strength of the system. We further develop a powerful method for a chosen plaintext attack showing non-fused transversal logarithmic signatures cannot be used. Moreover, proposed class of fused transversal logarithmic signatures withstand this attack when used in MST3 with Suzuki 2-groups and thus to our knowledge could be used to build secure realization of the scheme. We describe and discuss the implementation issues of the system in detail and include data of its performance obtained from an experimental result. Apart from the main research objective, we introduce a new approach to designing pseudorandom number generators based on random covers of finite groups. PRNGs based on random covers, called MSTg, turn out to be highly efficient for a certain class of group and produces high-quality random bit sequences. A very extensive sequence of tests for randomness using the NIST Statistical Test Suite and Diehard Battery of Tests provided here show extremely strong properties for the new methodology. More importantly, we show evidence that this class of generators is suitable for cryptographic applications. Finally, we include performance data of the generators and propose a method of using them in practice. [1] W. Diffie and M. E. Hellman, New Directions in Cryptography, IEEE Trans. on Inform. Theory, IT-22(6) (1976), 644–654. [2] S. S. Magliveras, B. A. Oberg and A. J. Surkan, A New Random Number Generator from Permutation Groups, In Rend. del Sem. Matemat. e Fis. di Milano, LIV (1984), 203–223. [3] S. S. Magliveras, A cryptosystem from logarithmic signatures of finite groups, in Proceedings of the 29’th Midwest Symposium on Circuits and Systems, Elsevier Publ. Co. (1986), 972–975. [4] S. S. Magliveras and N.D. Memon, Properties of Cryptosystem PGM, Advances in Cryptology, Lecture Notes in Comp. Sc., Springer-Verlag, 435 (1989), 447–460. [5] S. S. Magliveras and N.D. Memon, Random Permutations from Logarithmic Signatures, Computing in the 90’s, First Great Lakes Comp. Sc. Conf., Lecture Notes in Computer Science, Springer-Verlag, 507 (1989), 91–97. [6] S. S. Magliveras, Tran van Trung and D.R. Stinson, New approaches to designing public key cryptosystems using one-way functions and trap-doors in finite groups, J. of Cryptology, 15 (2002), 285–297. [7] W. Lempken, S. S. Magliveras, Tran van Trung and W. Wei, A public key cryptosystem based on non-abelian finite groups, J. of Cryptology, 22 (2009), 62–74.