From: aba@dcs.exeter.ac.uk
Date: Mon, 17 Jul 95 13:41:49 PDT
To: hayden@edu.msus.mankato.krypton
Subject: Re: RC4 crack
In-Reply-To: <Pine.ULT.3.91.950717151714.25405B-100000@krypton.mankato.msus.edu>
Message-ID: <20191.9507172036@exe.dcs.exeter.ac.uk>
MIME-Version: 1.0
Content-Type: text/plain



> I remember when RSA129 was being done, the program you have you manually 
> get a start location, and then email transparent any results that it 
> got.  The program that doled out areas to search would base those on what 
> had already been mailed in.  I don't know the details of how exactly that 
> worked, however.

Yeah it's quite like that except we're going for sockets, and an SMTP
style protocol.  That way people can write other apps to the protocol,
for instance Andy Brown has an SSL bruter and key management s/w for
NT, and he plans to interface to the 'master' software via this socket
protocol, allows intermixing, so some people will be running direct
IP, others with PCs or behind firewalls will be running via the WWW
interface which also talks the SMTP style stuff to the master, and it
would be possible if desired to write an email gateway to the socket
protocol for interacting with the master.

Also the socket protocol (blame Piete for this clever stuff, and most
of the socket protocol design) is planned to work with arbitrary
levels of masters, so you can start a local master say on your local
network, the local master requests keys of the 'big master', and doles
them out to 'slaves' running on each cpu you have.  When all it's
slaves have acked the keyspace it has drawn out from the big master,
it'll ack that bigger keyspace with the bigmaster and draw out some
more keyspace.

> But, if the program could be written in such a way that it was all 
> automatic, mailing in results and automatically (maybe via a telnet 
> port?) getting the information about what to search, that would be most 
> nice. 

Yep a telnet port is it for both reporting and getting keys, also the
WWW interface to the same.

> I'd basicly like to be able to start the program, nice it, slam it
> in the background, and forget about it.

Right, niceing seems to be one option another is to suspend it whilst
people are directly logged in, Kevin and some others have tools for
this kind of thing.  Also there was a similar ultra-nice batch job
suspender which came with RSA129, which we might pinch/combine.

The problem with nicing is that most unix schedulers don't seem to
know what nice means,.. you still get a noticable slow down on
interactive jobs on SGI boxes even if you've got it npri -h 150, and
even though the bruterc4 (and the bruteSSL too) have tiny resident
core sizes).  Also we thought there should be an hours of play option
so you can tell it (the slave) when it is allowed to hammer the
machine, say 6pm - 7am or whatever.

So, yes the idea that you can slam it in the background and forget it
is a very nice one as it ensures max resource usage.  Also it would
allow us to setup a semi-permanent key cracking ring, with slaves that
can support cracking both SSL and RC4, plus whatever anyone else adds
later, you would get to install a new "ability" then your machine
would say know how to do relations for a RSA-512bit or whatever.

Interesting to see how many MIPs can be mustered en masse for this
kind of app.

Adam