From: "     Toth"                         <1b2herma@nun.iut2-grenoble.fr>
Date: Thu, 30 May 1996 04:03:30 +0800
To: cypherpunks@toad.com
Subject: Re: Re: RSA breaking
Message-ID: <199605291505.RAA102870@nun.iut2-grenoble.fr>
MIME-Version: 1.0
Content-Type: text/plain


>        System of Russian dolls? What would you suggest?
  My idea(it is not me that found that) is that tou could have many public keys, when someone want to send you something, he just mail you his(or her) keys, and then you encrypt with the first key, then with the second, etc...
  This system is used by some anonymous remailers:
Each one has a key, you encript your text with the key of the last remailer, then you put your text, the adress of the people tou want the letter to be sent, then encrypt it with the key of the before the last remailer, and put the adress to be sent of the last remailer, etc
So your letter is decripted by the first remailer, then sent to the adress found, then the second remailer decrypts it, etc, like an onion.
  The fact is that you can obtain a very good privacy if you have some keymaker programm. Let suppose you want to send me a crypted text(don't, it's forbidden to crypt in france), you mail to my remailer, that choose 1000(arbitrary number) keys, that send it to you, crypted with your key, you crypt your text the necessary number of times.
  This way, the message will be maybe uncrypted by someone, but if the key change for each message, someone can't find your key, i insist on this danger, having one key is useless as soon as someone found it.

>What kind of system do you think would be needed to crack RSA coded messages?
  As i said, it is impossible to break an RSA code with softwares, the best attempt was 25 days with ten Crays linked for a key of only 50 numbers.
  But this can be done as i said before with a simple eratosten crible "translated" into hardware. Sorry electronic is not my way, but for the exemples i saw in class, it is really to the ability of a basic electronician, and they were more complicated than what i am talking about.
  The best components you have, the faster you'll go, so a governement or a bog private lab can afford this kind of prices.