1997-09-25 - Re: The CipherSaber Manifesto

Header Data

From: Antonomasia <ant@notatla.demon.co.uk>
To: reinhold@world.std.com
Message Hash: 1fdd0733f504ba4c2cc47f2a993977c6c6eb5c8f1119d2f5ea44dbbebd55702e
Message ID: <199709250042.BAA08154@notatla.demon.co.uk>
Reply To: N/A
UTC Datetime: 1997-09-25 09:03:01 UTC
Raw Date: Thu, 25 Sep 1997 17:03:01 +0800

Raw message

From: Antonomasia <ant@notatla.demon.co.uk>
Date: Thu, 25 Sep 1997 17:03:01 +0800
To: reinhold@world.std.com
Subject: Re: The CipherSaber Manifesto
Message-ID: <199709250042.BAA08154@notatla.demon.co.uk>
MIME-Version: 1.0
Content-Type: text/plain



reinhold@world.std.com (Arnold Reinhold) wrote:

> CipherSaber-1 (CS1) uses Ron Rivest's RC4 algorithm as published in
> the second edition of Bruce Schneier's Applied Cryptography. ....

> CipherSaber-1 is a symmetric-key file encryption system. Messaging
> takes place by attaching binary files to e-mail. Because CipherSaber
> uses a stream cipher, an initialization vector must be used to prevent
> the same cipher key from being used twice. In encrypted CipherSaber-1
> files, a ten byte initialization vector precedes the coded data. For
> decryption, the initialization vector is read from the file and
> appended to the user key before the key setup step.  ......

Why not _prepend_ the IV to the key ?  As described here any
paranoics who use keys > 255 chars won't get the IV in place, and
will lose out.  I think I'd also force 4 bytes of the IV to be the
current time, as a defence against the (P?)RNG getting me a repeated IV
eventually.

--
##############################################################
# Antonomasia   ant@notatla.demon.co.uk                      #
# See http://www.notatla.demon.co.uk/                        #
##############################################################






Thread