Re: Cipher "arikahs"

But if you really want to know, yes it is faster than AES.... so now
you know
No it is not smaller, actually adress 500 byte= 2*256 for the
permutation arrays.
And the keys ( key[], invkey[], savestate[], blockkey[] ) each 32
byte=256 bit.
And then you need one buffer variable 1 byte, so it is not smaller you
probably would like 1K memory to run, a minimum of 629 byte memory in
block mode.

Yes it is alot more safe than Rijndael due to the rekey of the internal
state, it make traditional cryptanalyse worthless when the actual
metablock/rekeyed permutation expansion take place.
You can chose size in multiples of 256 bit for these rekeyed and
expanded permutations.
If you chose a metablock of 1K rekeyed expansion there will not be any
material to analyse.

So yes it is alot safer than the AES candidate Rijndael, and there is
no need to ever change key, since the rekey algorithm let any password
hash/morph into all possible positions.

JT

Tom St Denis skrev:

jt64@xxxxxxxx wrote:
A cipher built from my automatic keyrescheduler.
The cipher totally byte scalable, due built around permutations of
bytes. That swap
Assume original key 256 bit

<snip gibberish>

Have you cryptanalyzed this? What attacks have you proven don't work
on it?

Have you benched this? Is it faster than AES? Smaller? More
efficient?

In short, why do YOU think this is a good idea?

Tom

.

Quantcast