Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

AES decryption (ASIC)

0 views
Skip to first unread message

swissi...@gmail.com

unread,
Aug 26, 2008, 3:51:56 AM8/26/08
to
Hey folks , i need ur opinion about something :
To implement an AES decryption (CBC mode ) algorithm in ASIC , what
would be the best way to do it ? i mean among these architectures
which one do you choose and why :

* Basic iterative architecture
* Partial loop unrolling
* full loop unrolling
* Partia outer-round pipelining
* Full outerround pipelining
* Inner-round pipelining
* Partial mixed innerand outerroundpipelining
* Full mixed inner- and outer-round pipelining
* other ...

Jon Beniston

unread,
Aug 26, 2008, 5:33:46 AM8/26/08
to

What process are you using and what are your power / area / bandwidth
requirements.

Jon

Message has been deleted

dajjou

unread,
Aug 26, 2008, 7:42:58 AM8/26/08
to

Thank you jon ,
In fact i need to know the architecture of such algorithm ( AES 256
decryption CBC mode) inside an ASIC that i don't know anything about
it (area ,bandwith ...).
my question is how many possibilities to implement this algorithm ?
could we fix all possibilities ?

Jon Beniston

unread,
Aug 26, 2008, 8:35:51 AM8/26/08
to

> Thank you jon ,
>  In fact i need to know the architecture of such algorithm ( AES 256
> decryption CBC mode) inside an ASIC that i don't know anything about
> it (area ,bandwith ...).
> my question is how many possibilities to implement this algorithm ?

You can implement AES in many different ways. Which is right for you
will depend upon what your requirements are.

Jon

arsonper...@gmail.com

unread,
Aug 26, 2008, 9:58:46 AM8/26/08
to

Keep in that you'll only need to do the roundkey expansion once for
each chain of CBC operations, so if you can compute that and store it
in a register file of some sort. If you already have existing AES
encryption IP, be sure to look at 5.3.5 of FIPS-197. You may be able
to implement the decryption operation by simply changing the internal
blocks to their inverse functions.

Mike Treseler

unread,
Aug 26, 2008, 11:34:34 AM8/26/08
to


Is this multiple choice or an essay question?

Will this be on the test?

Thomas Stanka

unread,
Aug 27, 2008, 1:55:02 AM8/27/08
to

the first is the smallest and slowest. You choose it, when small area
is more important than throughput. I'm no AES-Expert, but full loop
unrolling means you need ~14 times the area and gain a bit less then
14 times throughput.

Ask your self what you need.

bye Thomas

0 new messages