Rob Shirey writes:
I got Ray Lau's RIPEM Mac 0.7b1 for my Mac IIci. In signing and
enciphering short messages, using a "512-bit key size", it seemed to
operate almost instantaneously. So I took a ~250K byte Word file, and
uuencoded it, which made it about ~740K bytes. Signed-only
(MIC-Clear) >>took 20 wallclock seconds. Signed and enciphered
(ENCRYPTED) took about 35 >>seconds. To me, this seems fast enough to
do real work!
RIPEM is based on RSAREF. Has anyone on the list compared their own
crypto >>or a BSAFE-based verstion to an RSAREF-based version on the
same machine?
Ah, a challenge! I just encrypted and signed a 270KB file and using
SecretAgent/Mac 1.0.4. This did the following:
1. computation of pseudorandom DES key based on 20-byte random seed
and number of ticks since last boot (uses NIST SHA for state
transitions; use of a purely random key generation scheme is
optional)
2. lookup of recipient's public key; ElGamal encryption of DES key
3. DES encryption of file (270KB) in CBC mode
4. NIST Secure Hash of file (preformed simultaneously with 2) as
message digest
5. lookup of sender's public key, validation of private key, and
computation of NIST DSA on hash
Took about 25 seconds. (I'm running a IIci also, but it's got a 50MHz
68030 accelerator and 32MB RAM installed with a large disk cache; you
don't say what you're running.) To be fair, I should also point out
that no printable encoding was done in my case--I assume RIPEM does
that--which probably explains why my time is slightly better. Just
kidding, I'm sure Lau's code is considerably better than our C++/MacApp
thing, but it's nice to see we're in the same ballpark. :-)
Sorry to see you still messing with that RSA stuff--I thought you were a
convert. :-)
mjm
----------
Michael J. Markowitz, VP R&D
markowitz(_at_)dockmaster(_dot_)ncsc(_dot_)mil
Information Security Corp. 708 405-0500, fax: 708 405-0506
1141 Lake Cook Rd., Suite D MCI: 363-1959
Deerfield, IL 60302 CIS: 76206,2617