AES Candidates: A Survey of Implementations

We present a cross-table of almost all publicly known implementations of AES candidates, including the ones done by the authors. A short overview of our own implementations of Rijndael is given. The relative easiness of doing “the world best” implementations and a lot of gaps in the table force us to ask if there is enough information known to really decide which ciphers are fast and which are not. (This paper only compares the encryption speed in the case of the 128-bit keys.) In the conclusions we present a very brief survey of the known attacks to the candidates, stressing the fact that other aspects of the candidates are still less known (at least to the public). We finish this paper with apparent conclusions that the first round of the AES process has been too short, but we still give recommendations which candidates should be elected to the second round basing on the knowledge known to the public at the moment of writing this paper.