Re: A variation of RC4

From: Mok-Kong Shen (mok-kong.shen_at_t-online.de)
Date: 07/09/04 

Date: Fri, 09 Jul 2004 18:42:03 +0200

Bill Unruh wrote:
> Mok-Kong Shen <mok-kong.shen@t-online.de> writes:
>
> ]Addendum:
>
> ]RC's swapping of elements of an S-box to render it variable
> ]dynamically could be considered to be in a sense in the same
> ]vein as swapping of regions of elements of the array that is
> ]done in Schneier's Solitaire, which results in correspondingly
> ]larger changes of the S-box. With the counters i and j of
> ]the original RC4, one has three regions as follow:
>
> ] 1 2 3
> ] ---------------------------
> ] i j
>
> ](In this constellation one counts S_i to region 2 and S_j
> ]to region 3. In case i=j, we increase j by 1.) One could
> ]then, similarly to Solitaire, swap region 1 with region 3.
> ]Since it seems to be advantageous operationally to have
> ]spaces of two arrays to facilitate copying and alternatingly
> ]use one of them as the current S-box, one could also easily
> ]reverse the order of the elements in the regions thereby,
> ]if desired.
>
>
> The advantage of RC4 is its speed. Slowing it down by a factor ot 100 does
> not seem a great advantage.Also, in general regions 1 and 3 are of
> different size.

One should better consider my description as indicating a
'general' possibility of dynamically modifying a given S-box
(which is not necessarily one for RC4), though the idea came
to me in the context of RC4. Thus constraint of speed may
(or may not) dictate whether it be employed in any given
situation. On the other hand, that larger modification step
could also be introduced only after every m original RC4
steps (for some chosen number m), thus correspondingly
reducing the speed impact. That regions 1 and 3 are of
different size is true in the general case. This is also
the case for Solitaire and should impose no problem, the
consequence being only that the region 2 gets shifted in
position.

M. K. Shen

Relevant Pages

  • Re: 8 Bit Random Numbers
    ... RAM, I would echo nospam's suggestion to use RC4 (aka ArcFour, ... A description of ARCFOUR (Alleged RC4), written by Neil Bawd in 1997 ... Every time the array is used it changes - by swapping two bytes. ...
    (sci.electronics.basics)
  • Re: doubts about rc4
    ... >for rc4. ... the number of cycles are varying as function ... >realized the array has to be filled with the key size and, in turn, ... Greg Rose ...
    (sci.crypt)
  • Re: Securing ARC4
    ... Is this applied before the normal RC4 ... Say you generate one S-table with key K and array ... for the attacker to choose or control R in some way. ... Crypto primitives are hard to evaluate. ...
    (sci.crypt)
  • Re: Securing ARC4
    ... Is this applied before the normal RC4 ... Say you generate one S-table with key K and array ... for the attacker to choose or control R in some way. ... Crypto primitives are hard to evaluate. ...
    (sci.crypt)