Investigations in the Design and Analysis of Key-Stream Generators

The work is focused on several important problems in
the design and analysis of key-stream generators for
stream ciphers. We look at common building blocks
that are used in key-stream generators. These are
nonlinear logical functions and clock-controlled
linear feedback shift registers (LFSR). We introduce
tensor transform which gives new tools for analyzing
security of the functions. We also study correlation
properties of Boolean functions. A newly introduced
probabilistic function of a Boolean function
estimates the probabilistic distribution of its
output bits if the distribution of the arguments,
the function depends on, is known. For a clock-
controlled LFSR, we estimate the period of its
output sequence when the feedback polynomial is
irreducible and the structure of the control
sequence is arbitrary. Away from building blocks, we
construct a key-stream generator based on the one
suggested by Geffe. Unlike the Geffe generator that
has three binary input m-sequences, this generator
runs over the extension field and combines multiple
inputs having arbitrary periods. Finally, we develop
several statistical attacks on stream ciphers.

Autorentext
A.Kholosha received Ph.D. from the Eindhoven University of Technology, The Netherlands, in 2003. Currently, researcher at the Department of Informatics, University of Bergen, Norway. Research interests: cryptographic properties of Boolean and pseudo-Boolean functions, sequences and key-stream generation for stream ciphers, cryptanalysis.

Klappentext
The work is focused on several important problems in the design and analysis of key-stream generators for stream ciphers. We look at common building blocks that are used in key-stream generators. These are nonlinear logical functions and clock-controlled linear feedback shift registers (LFSR). We introduce tensor transform which gives new tools for analyzing security of the functions. We also study correlation properties of Boolean functions. A newly introduced probabilistic function of a Boolean function estimates the probabilistic distribution of its output bits if the distribution of the arguments, the function depends on, is known. For a clock- controlled LFSR, we estimate the period of its output sequence when the feedback polynomial is irreducible and the structure of the control sequence is arbitrary. Away from building blocks, we construct a key-stream generator based on the one suggested by Geffe. Unlike the Geffe generator that has three binary input m-sequences, this generator runs over the extension field and combines multiple inputs having arbitrary periods. Finally, we develop several statistical attacks on stream ciphers.