[601] in cryptography@c2.net mail archive
RPK?
daemon@ATHENA.MIT.EDU (Keith Dawson)
Mon Apr 21 13:27:57 1997
Date: Sat, 19 Apr 1997 19:14:37 -0400
To: cryptography@c2.net
From: dawson@world.std.com (Keith Dawson)
Any opinion in this community on William M. Raike's RPK public key
cryptosystem? I checked with Antonomasia's robot and it doesn't seem
to have been discussed here. The tech info page says:
"As its 'one-way function,' RPK uses discrete exponentiation over
multiple finite fields GF[2^n]. It shares these mathematical
foundations with other cryptosystems, notably the Diffie-Hellman key
distribution system and the El Gamal cryptosystem, and as a result
gains part of its security from the fact that the best known attack
involves computing discrete logarithms, a process whose difficulty is
considerable and quantifiable."
RPK's pages claim that the algorithm is particularly well suited to
efficient implementation in silicon. The performance of a software
implementation is no slouch either; again from the tech info page:
"On a 16MB Pentium 90, the RPKFILE evaluation software running under
Microsoft Windows 95 with the default security settings (737 bit keys)
requires less than 2 seconds to create a public key. Encryption of a
1MB file requires about 13 seconds the first time a public key is used
and about 8.5 seconds for subsequent encryptions with the same public
key. Decryption takes about 12 seconds."
Public and private keys do not commute so the system is not suitable
for signatures and authentication in the same way RSA is.
Home
<URL:http://www.rpk.co.nz/>
Concept
<URL:http://www.rpk.co.nz/concept.htm>
Tech info
<URL:http://www.rpk.co.nz/techinfo.htm>
Security comparison with RSA
<URL:http://www.rpk.co.nz/secrmark.htm>
SafeCracker Challenge
<URL:http://www.rpk.co.nz/reward.htm>
