A new aspect of dual basis for efficient field arithmetic

In this paper we consider a special type of dual basis for finite fields, GF(2^m), where the variants of m are presented in section 2. We introduce our field representing method for efficient field arithmetic(such as field multiplication and field inversion). It reveals a very effective role for both software and hardware(VLSI) implementations, but the aspect of hardware design of its structure is out of this manuscript and so, here, we deal only the case of its software implementation(the efficiency of hardware implementation is appeared in another article submitted to IEEE Transactions on Computers). A brief description of several advantageous characteristics of our method is that (1) the field multiplication in GF(2^m) can be constructed only by m + 1 vector rotations and the same amount of vector XOR operations, (2) there is required no additional work load such as basis changing(from standard to dual basis or from dual basis to standard basis as the conventional dual based arithmetic does), (3) the field squaring is only bit-by-bit permutation and it has a good regularity for its implementation, and (4) the field inversion process is available to both cases of its implementation using Fermat’s Theorem and using almost inverse algorithm[14], especially the case of using the almost inverse algorithm has an additional advantage in find- ing(computing) its complete inverse element(i.e., there is required no pre-computed table of the values, x^-k, k = 1, 2,…).