From: andreas@artcom.de (Andreas Bogk)
Date: Thu, 22 Feb 1996 20:01:51 +0800
To: IPG Sales <ipgsales@cyberstation.net>
Subject: Re: Internet Privacy Guaranteed ad (POTP Jr.)
In-Reply-To: <Pine.BSD/.3.91.960221145534.3814H-100000@citrine.cyberstation.net>
Message-ID: <y8aybpww62l.fsf@horten.artcom.de>
MIME-Version: 1.0
Content-Type: text/plain


-----BEGIN PGP SIGNED MESSAGE-----

>>>>> "IPG" == IPG Sales <ipgsales@cyberstation.net> writes:

    IPG> I find less and less disagreeement with your comments - with
    IPG> one major exception - for a given message length - say 10 to
    IPG> the 500th power, a OTP seeded algorithm, a better term would
    IPG> be to call it an OTP driven algorithm, can produce the exact
    IPG> same effect as an OTP of that length - that is, the encrypted
    IPG> text can be any possible message of that length, and it is
    IPG> not possible to predict in way what the RNG generated stream
    IPG> is -

First, what you describe is commonly called a keyed RNG. Such a system
is provably less secure than an OTP, because the number of possible
plaintexts from any given ciphertext is limited by the number of
possible keys. This makes an exhaustive search of all keys possible,
because it is very unlikely that a given ciphertext decrypts to
multiple plaintexts that make sense. In contrast, with a OTP there are
as many keys as there are possible plaintexts, so any given plaintext
can be reached, making it impossible to recognize the correct
plaintext. Of course, searching the whole keyspace might be impossible
if the number of possible keys is large enough.

But there are other ways of attacking a croyptosystem besides trying
all possible keys. Your cryptosystem seems to be based on what is
called a linear congruential generator in combination whith an
RC4-like 8*8 S-box, although somewhat simpler.

I don't want to make any claim about the security of the algorithm,
but linear congruential generators can't be considered secure for any
cryptographic use. Your only chance is that the security of that
algorithm does not depend on the generator, but I doubt that.

For further reference, go out and buy "Applied Cryptography" by Bruce
Schneier.

The pseodo-code snipped describing your algorithm, for other people's
reference:

    IPG> Bi=(Bi+Ci MOD Di) Mod 256              Large prime numbers
    IPG> ENCRYPTEXTi=OTP[Bi] XOR PLAINTEXTi     Encryption
    IPG> OTP[Bi]=ENCRYPTEXTi                    Makes the OTP Dynamic

Andreas

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2
Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface

iQCVAgUBMSu6OEyjTSyISdw9AQGKIQP+MqE5Scq99kGfLT8CdN3h9abJZNhj9qzm
rUFGsnXfdAvyRzfLz6v8FsfLHgnkgu10MG++NABFBz0I+U0iGFi8Zivkd3Ae9/6J
qOHqbGjiS4r3QN8IOLDwAW6eO6pF4Z0A/+FqLVR+zB+OZF/7TzUmgWpa8+cLWQkH
Hndr5tAVekw=
=bY+f
-----END PGP SIGNATURE-----