Hi there,
On Sun, 25 Jun 2000, Wyman Eric Miles wrote:
> I can give a counterrecommendation, at least. We have a Cryptoswift 100
> PCI demo unit. I recompiled OpenSSL 0.9.5a with the patches Rainbow
> supplied.
>
> The first card we got was DOA. The system (Ultra 5/Solaris 2.6)
> recognized the 2nd fine. For a while. The OpenSSL speed benchmarks would
> hang and applications which required RSA keys were a little flighty. We
> never did get an adequate, verifiable measure of performance from the CS
> 100. After a day or two, the kernel module for the card lost track of the
> thing and no amount of effort since has allowed us to re-establish rapport
> with the device.
I won't get into discussing the merits (or otherwise) of the various
different hardware manufacturers, but will simply mention that I've been
running a CryptoSwift card in a test setup for some time, periodically
with some heavy load testing, and haven't had any problems. I'm running a
Linux setup BTW. Yes, YMMV but so far my mileage has been pretty good.
On a related subject, if you are interested and able, the "BRANCH_engine"
branch of the OpenSSL CVS repository has some work in progress providing
hardware support for various cards; CryptoSwift, Atalla, and nCipher are
in there so far to varying degress of completeness ... feel free to have a
play if you're interested in it.
Cheers,
Geoff
______________________________________________________________________
OpenSSL Project http://www.openssl.org
User Support Mailing List [EMAIL PROTECTED]
Automated List Manager [EMAIL PROTECTED]
