home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!pipex!bnr.co.uk!uknet!acorn!eoe!ahaley
- From: ahaley@eoe.co.uk (Andrew Haley)
- Newsgroups: sci.crypt
- Subject: Re: Anyone know about FEAL-8
- Message-ID: <1487@eouk18.eoe.co.uk>
- Date: 23 Dec 92 14:56:37 GMT
- References: <1992Dec22.134951.16838@ncsu.edu>
- Organization: EO Europe Limited, Cambridge, UK
- Lines: 42
- X-Newsreader: TIN [version 1.1 PL6]
-
- JAMES LEWIS NANCE (jlnance@eos.ncsu.edu) wrote:
- :
- : I have been reading a 1988 paper by Miyaguchi, Shiraishi, and Shimizu in which
- : they describe an encryption algorithm they call FEAL-8. I am interested in
- : this algorithm because it was designed to run quickly on computers. I have
- : never heard of it anywhere before and I have the following questions about it.
- : Can anyone help me?
- :
- : 1) Is FEAL-8 Secure (as compared to DES)?
-
- It's not as secure as DES. There is a chosen plaintext attack which
- can crack FEAL-8 (when used in ECB mode) with a couple of thousand
- encryptions. This does require, however, that the FEAL-8 user encrypt
- these plaintexts sent to him by an attacker. This attack isn't a
- problem if:
-
- a. You change keys frequently, by using a key exchange method such as
- Diffie-Hellman or RSA.
- b. You never use ECB mode, but use CBC instead.
-
- FEAL-32 (FEAL with 32 rounds) appers to be as resistant to
- cryptanalysis as DES, at least for the time being.
-
- : 2) Is C source code available for it anywhere?
-
- There's no need. The algorithm is trivial, and any competent C
- programmer should be able to code it in a few hours. I recommend the
- use of assembly for high speed implementations.
-
- : 3) Is it patented
-
- Not as far as I know.
-
- : 4) Is it a good choice for a fast, secure private key method?
-
- Probably not, because of the vulnerability to attack mentioned in the
- answer to question 1. IDEA is an alternative, but is patented. I
- suggest you also look as Khufu, which was described recently in one of
- the CRYPTO proceedings (I think '91?). This is a cipher from Xerox
- PARC, designed by Ralph Merkle. I think it's patented too.
-
- Andrew.
-