Computation of Edit Probabilities and Edit Distances for the A5-Type Keystream Generator

Abstract

Several edit probabilities and edit distances incorporating the stop/go clocking in the well-known A5-type keystream generator are proposed. Recursive algorithms for their efficient computation are derived. It is shown how the edit probabilities can be used to mount statistically optimal correlation attacks on any two of the three stop/go clocked shift registers. By using a method for estimating the underlying false alarm probability, it is argued that the minimum output sequence length required to be known for a successful attack is linear in the total length of the respective shift registers.