New Approach in the Rainbow Tables Method for Human-Like Passwords

This paper represents a new approach to rainbow tables, a method of password recovery that was originally developed by Martin E. Hellman and then improved by P. Oechslin, so most of its implementations use Oechslin's modification. An improvement represented in this work mostly lies in the reduction function, which uses character statistics to generate more "human-like" passwords. Though it generates passwords 5 to 10 times slower than reduction function, which uses direct dependency between hash bytes and the inserted characters, it significantly increases common efficiency in memory (8 to 30 times less memory needed to store these tables) and successful "human-like" passwords recovery probability, while these tables are generated by the same time as tables with the use of "random" reduction function.