OTP seed solution? - strong, tried before???

A very simple idea came to me today that I`m sure has been done 
before and I wanted to find out if it has any problems I haven`t 
seen:

A strong random generator (ie. a BBS) is seeded with a true random 
seed (derived possibly from keyboard latency) and used each time a 
message is send to create a message length randon string.

This string is XOR`d or added to the message creating a OTP.

The recipient has previously been sent a seed value for the generator 
encrypted under say RSA and signed to prevent a man in the middle 
attack.

Each set of recipient<->sender pairs has a seed unique to them.

This seems good to me but the key distribution mess (ie. hundreds of 
keys about) is the big problem, has anyone done any research into 
possible solutions to this???


  Datacomms Technologies web authoring and data security
     Paul Bradley, Paul@fatmans.demon.co.uk
       Http://www.fatmans.demon.co.uk/crypt/
         "Don`t forget to mount a scratch monkey"

-----BEGIN PGP PUBLIC KEY BLOCK-----
Version: 2.6.3ia
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=CMZA
-----END PGP PUBLIC KEY BLOCK-----

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

On Tue, 17 Sep 1996 paul@fatmans.demon.co.uk wrote:

> A very simple idea came to me today that I`m sure has been done 
> before and I wanted to find out if it has any problems I haven`t 
> seen:
> 
> A strong random generator (ie. a BBS) is seeded with a true random 
> seed (derived possibly from keyboard latency) and used each time a 
> message is send to create a message length randon string.
> 
> This string is XOR`d or added to the message creating a OTP.
> 
> The recipient has previously been sent a seed value for the generator 
> encrypted under say RSA and signed to prevent a man in the middle 
> attack.

That's the definition of a stream cipher.  BBS is rather slow and is dependant
on the difficulty of factoring large numbers.  Of course, if the keys are
distributed using RSA or Rabin, then the strength of the system depends on the
difficulty of factoring numbers anyway.  There are many stream ciphers that
are fast and secure.  I generally like to use a block cipher in CFB or OFB
mode rather than a stream cipher, but they are basically the same thing.

Mark
- -- 
PGP encrypted mail prefered.
Key fingerprint = d61734f2800486ae6f79bfeb70f95348
http://www.voicenet.com/~markm/

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3
Charset: noconv

iQEVAwUBMkNA2CzIPc7jvyFpAQHeZAf/d6cdUZ0611N+9E5yTwLebrJtdMJJG0zo
jk1rujKQB/3+Faexrazw2hTjTdoidH/irSWrCSPlgdgPOO8kuwK5HSONlXD/gzK3
WP0lS5iiTHSg0Gfge/K2px4HJjl4gP/bsNkcdTDf5QuZHsJFQ1wKcBkuv5AVCuD8
wp3fGKBy2fD2HrAAbvmACHwzVYV99D6X7KIXkBUB8kvH4cBPGwe1dWH45uWtE5UW
UJRyxax2m31K7mJ5kcIEf/noQHikZVCBNx0ojlnMub2M+UZhenJvvoVMu/0AqF7g
1OCk7H7SEI5g9AzB/zboKh+6Zs5oII+lsIIdNZRz4xqyaNzmQUU/Lg==
=+n1m
-----END PGP SIGNATURE-----