john-users - Re: what is/was wrong with dnetj?

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Wed, 23 May 2012 20:43:12 +0400
From: Solar Designer <solar@...nwall.com>
To: john-users@...ts.openwall.com
Subject: Re: what is/was wrong with dnetj?

Aleksey -

On Mon, May 14, 2012 at 10:31:35PM +0400, Aleksey Cherepanov wrote:
> I looked through http://openwall.info/wiki/john/parallelization and
> got a question: what is wrong with dnetj for distribution? Why is not
> it popular?
> 
> It seems that dnetj is intended to be what is needed for distributed
> attacks like during the contest, is not it? What makes it far from its
> intentions?

Frankly, I never tried out dnetj.  I only looked at the code, found it
reminiscent of my johnet experiment from 1997, but also unfinished and
IIRC using a less optimal way to distribute portions of incremental
mode's keyspace.  If I were to work on implementing distributed
processing for JtR now, I'd start from scratch anyway - using only
experience gained from past experiments, but not any code.

So maybe dnetj was/is in fact fine for some use.  Its actual code was
just not useful for a potential distributed processing implementation
that I could start the work on.

Maybe you can try it out and report to john-users?

If you want to review the code, I think dnetj was cloned off JtR 1.7.2,
so that would be the version to diff dnetj against.  You may download
old versions of JtR from:

http://download.openwall.net/pub/projects/john/

Thanks,

Alexander