In this paper, we revisit the recent small characteristic discrete logarithm algorithms. We show that a simplified description of the algorithm, together with some additional ideas, permits to obtain an improved complexity for the polynomial time precomputation that arises during the discrete logarithm computation. With our new improvements, this is reduced to O(q6), where q is the cardinality of the basefield we are considering. This should be compared to the best currently documented complexity for this part, namely O(q7). With our simplified setting, the complexity of the precomputation in the general case becomes similar to the complexity known for Kummer (or twisted Kummer) extensions.
CITATION STYLE
Joux, A., & Pierrot, C. (2014). Improving the polynomial time precomputation of frobenius representation discrete logarithm algorithms: Simplified setting for small characteristic finite fields. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8873, pp. 378–397). Springer Verlag. https://doi.org/10.1007/978-3-662-45611-8_20
Mendeley helps you to discover research relevant for your work.