From: aba@dcs.exeter.ac.uk
Date: Fri, 18 Aug 95 09:08:43 PDT
To: dan@milliways.org (Dan Bailey)
Subject: Re: Where is the key cracking farming software?
Message-ID: <14470.9508181605@exe.dcs.exeter.ac.uk>
MIME-Version: 1.0
Content-Type: text/plain



Dan Bailey <dan@milliways.org> writes on cpunks:
> There must be several versions of the code at this point.  The
> Cypherpunks release, Damien's release and whoever else wrote some code
> to do the chore.  If someone could pass me a pointer to the version
> that's best-commented and most understandable (one man's C is another
> man's crypto) I'd like to port it to Windows NT and write a simple
> installation to install it as a service.

Yep there's lots of versions of the brutessl software, ones I know of
(in no particular order):

- Damien Doligez wrote one
- Andrew Roos wrote on (this is the one we're using for the challenge)
- Eric Young wrote one
- I wrote one

More related softwares, specifications etc.

- Piete Brookes, Andy Brown, and I wrote a protocol specification for
  a SMTP style key doler - SKSP (Simple Key Search Protocol) - draft
  RFC like document on brute (www url below).

- Piete wrote a unix socket based key server and client for generic
  unix machines (in perl).

- Andy Brown wrote a Windows NT client which talks the same protocol

- David Byers did a MasPar port of the brutessl code

> 	Perhaps a general-purpose OO bruteforcing library (addition to
> Crypto++?) would be a good idea.  Just pass a pointers to the
> encrypt() and decrypt() functions to use for this session to the
> Cracker object, which in turn takes a number of bits and starts
> cracking.:)

The socket based key distributer allows you to write clients which
interact with the key server.  Take a look at Andy's code.

> 	The problem with this is the disparity among out-of-the-box encrypt
> and decrypt functions.  But I suppose it's nothing that couldn't be
> overcome with wrapper functions.  Just thinking out loud. :)

Not sure how Andy addresses this in the NT code (I don't have NT), but
for the unix client you'd just get another brute force program,
compile it and update the client.

Piete has a WWW URL for the software (the protocol spec is there too) at:

	http://www.brute.cl.cam.ac.uk/brute/

(Brute is a subsubdomain he set up for the purpose).

Also this:

	ftp://ftp.brute.cl.cam.ac.uk/pub/brute/

should work.

And the socket server runs on: sksp.brute.cl.cam.ac.uk (port 19957).

I think we're ready to start another challenge, this one a
collaborative project like the brute rc4 one.  Just need some
challenges now (asked Hal about this).

Adam