Re: (long) An AES implementation for 32-bit platforms
From: Mok-Kong Shen (mok-kong.shen_at_t-online.de)
Date: 06/11/03
- Next message: Richard Heathfield: "Re: (long) An AES implementation for 32-bit platforms"
- Previous message: Mok-Kong Shen: "Re: (long) An AES implementation for 32-bit platforms"
- In reply to: Mok-Kong Shen: "(long) An AES implementation for 32-bit platforms"
- Next in thread: Mok-Kong Shen: "Re: (long) An AES implementation for 32-bit platforms"
- Reply: Mok-Kong Shen: "Re: (long) An AES implementation for 32-bit platforms"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: Wed, 11 Jun 2003 00:10:57 +0200
Addendum: A few points of clarification:
(1) The copyright notice is so formulated so as to
gather informations of its practical use, in
particular matters concering possibilities of
efficiency enhancements or eventual bug reports,
to enable its further improvements in furture
(major) versions to be centrally distributed by
me to all who are interested in it (via this group).
(2) The code has been carefully (as far as I personally
could) written to be easily readable, instead of
employing by all means complicated coding techniques
for the purpose of squeezing out efficiencies. This
should facilitate visual checking of its logic and
renders its maintenance simple. (Also I don't use
conditional compiler directives or macros, which in
my personal experience are not very contributive to
easy understanding of codes.) I had in fact an
optimized version of one of the functions that is
needed in key expansions. Since I consider that one
likely changes keys relatively seldom in comparison
to processing of the blocks, I opted for using the
straightforward code for that function, so that those
readers who desire to have a proper understanding of
the code before venturing to use it wouldn't have
difficulties therefrom.
(3) Until recently, I never wanted to implement AES
and had also not known how many implementations
are freely available to all. I was surprised, as
I looked at page that provides a survey of
diverse implementations. This convinced me that
there may be a place for another implementation
that could be straightforwardly used by a large
number of people on the most porpular hardware
type at least. (I don't have a big-endian
hardware to test the code, but I am confident
that there shouldn't be problems arising from
endian-ness. 'Bit-sex' is a different issue,
which could eventually need conversion of the
individual bytes on external data exchange media
or equivalent measure. Cf. the topic of network
standard format.)
(4) As said, part of the initiative for doing the work
comes from discussions in another recent thread.
I wanted to actually see whether a straightforward
(possibly naive)interpretation of the FIPS document
is or isn't sufficient by itself for working out a
concrete byte-array implementation. (There were
opinions whether endian-ness of hardware could be
an issue in this context. In retrospect I believe
that that verification wouldn't be necessary. Anyway,
for one thing, AES is, as is well-known, also
intended for implementation on 8-bit processors and
there isn't an endian-ness issue (in the proper sense
i.e. concerning ordering of bytes in a word) on such
computers.)
(5) Anybody having difficulties in using my code but not
desiring to post the problems to the group could
e-mail me directly. (Should there be bugs found, I
would certainly inform the group.)
M. K. Shen
- Next message: Richard Heathfield: "Re: (long) An AES implementation for 32-bit platforms"
- Previous message: Mok-Kong Shen: "Re: (long) An AES implementation for 32-bit platforms"
- In reply to: Mok-Kong Shen: "(long) An AES implementation for 32-bit platforms"
- Next in thread: Mok-Kong Shen: "Re: (long) An AES implementation for 32-bit platforms"
- Reply: Mok-Kong Shen: "Re: (long) An AES implementation for 32-bit platforms"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|
