Another day I've read "CryptoSwift Helps Meet Client Demands" in June
issue of "Server/Workstation Expert." I have no idea why do I get this
magazine... In either case they wrote about Rainbow's "RSA in hardware"
card. They claim that it took 100% of CPU time to serve ~3 users on 167
MHz Ultra-1 and ~4 - on 200 MHz PPro with software RSA while with
Rainbow card it would took 20-30 users to saturate the CPU. I find the
first figures unfair. I'm *not* saying they cheat to make their product
look good. It's a great product featuring hardware random generator and
off-RAM storage for private keys. I simply believe that software (i.e.
OpenSSL) can do better than 3-4 users. No, I don't feel like doing what
they did to benchmark OpenSSL:-) There is another aspect I wonder about.
It's obvious that they've compared against RSAref library, right? So I
wonder if OpenSSL users who link with RSAref could run 'apps/openssl
speed rsa dsa' and submit the results (to me if you wish, I can
summarize) along with CPU description.

Andy.
______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       [EMAIL PROTECTED]
Automated List Manager                           [EMAIL PROTECTED]

Reply via email to