Re: RSA vs AES

From: LVB (LvB_at_sdc.net)
Date: 08/29/03 

Date: Fri, 29 Aug 2003 14:52:17 GMT

On Fri, 29 Aug 2003 06:06:45 -0700, Waldyr wrote:

> Hello !
>
> We know that symmetric algorithms are faster than assymmetric ones,
> for the same level of security. I'm interested to know if there is any
> reference where this difference is measured, I mean, about *how many
> times* AES-128, for example, is faster than RSA-1024, for the same
> plataform and the same volume of data.

        I have some data for IA64. For the same amount of data (about 128
bytes, which is the most you can do in one go with RSA-1024) AES-128 does it
in little more than 8 CPU cycles per byte, whereas a 1024-bit RSA private
key operation can do no better than 1980 CPU cycles per bytes. Most RSA
implementations are far less efficient than this though.