In eprint/2012/338 we read:

"for the first time, we propose a general algebraic attack framework on
the multi-output stream ciphers"

But in fact the authors ignore the most basic literature on the topic:
The following paper specifically covers the scenario with multiple
outputs and provides many useful worst-case bounds on the existence of
such attacks:

Nicolas Courtois:
Algebraic Attacks on Combiners with Memory and Several Outputs,
ICISC 2004, LNCS 3506, pp. 3-20, Springer 2005.
An extended and updated version of this paper is available at
eprint.iacr.org/2003/125/.

=============

On page 5 we find 4 attack scenarios called S1-S4.

This is highly confusing, because,
please note that a similar notation S12345 was used in the extended version of
Nicolas Courtois, Willi Meier: Algebraic Attacks on Stream Ciphers with
Linear Feedback. Eurocrypt 2003, LNCS 2656, pp. 345-359, Springer.
This is available at [www.nicolascourtois.com]

The notation S12345 is also used in extended slides by Courtois which can be found at
[www.nicolascourtois.com]

The scenario S5 on page 70 is precisely the attack with multiple outputs (and also for augmented functions).
See also:
[eprint.iacr.org]



Edited 7 time(s). Last edit at 10-Sep-2012 12:46 by ncourtois.

MOREOVER.
in their paper we read:

"This paper focuses on the LFSR-based multi-output stream
ciphers consisting of a linear feedback shift register (LFSR) and a multioutput filter boolean function"

WHILE my paper covers already a MUCH more general case when also you have internal memory added to the non-linear filter, and when you have 0 memory bits, my framework degrades to the simple case they study here.

The authors claimed to ignore and have promised me to update their paper on 25/06/2012...

However to this day their paper was NOT updated.

I must therefore ask the IACR to withdraw this paper and to blacklist the authors from eprint as perpetrators of serious scientific fraud.



Edited 3 time(s). Last edit at 10-Sep-2012 13:02 by ncourtois.