Other Directory Sites: SeekWonder | Directory Owners Forum
The following archive was created by hippie-mail 7.98617-22 on Sat Apr 10 1999 - 01:13:58
Mok-Kong Shen (mok-kong.shen@stud.uni-muenchen.de)
Wed, 02 Sep 1998 12:51:43 +0100
Martin Grap wrote:
>
> >On Mon, 31 Aug 1998, Sandy Harris wrote:
> >> But it looks to me like it isn't entirely hopeless, & even if it
> >> is, then proving it hopeless for a given cipher might be useful.
> If I remeber correctly, Shannon discussed this sort of algebraic
> cryptanalysis in his 1949 paper "A Communications Theory of Secrecy
> systems". As one possibility to thwart these attacks he recommends to
> use components which add a lot of "confusion" to the cipher. I.e.
> he recommends to use "complicated" non-linear mappings as building
> blocks for the cipher (today mostly in form of S-Boxes). Other building
> blocks should be used to "diffuse" the statistical properties of the
> plaintext in such a way that an enemy needs a large amount of intercepted
> material in order to be able to reconstruct these statistical properties.
> In other words these building blocks add diffusion to the cipher (today
> mostly in the form of permutations or special linear functions).
In view of the difficulties of solving non-linear equations as
compared to linear ones, I think that the use of purely algebraic
methods to attack any more or less reasonably designed encryption
algorithm of practical size is therefore computationally infeasible.
Perhaps new computing paradigma like quantum or DNA computing could
change this fact. I am skeptical even in that.
M. K. Shen
The following archive was created by hippie-mail 7.98617-22 on Sat Apr 10 1999 - 01:13:58