From: Andy Brown <a.brown@nexor.co.uk>
Date: Wed, 24 Jul 1996 00:28:16 +0800
To: "'Ben Holiday'" <ncognito@gate.net>
Subject: RE: Distributed DES crack
Message-ID: <01BB787C.91653EF0@mirage.nexor.co.uk>
MIME-Version: 1.0
Content-Type: text/plain


On 22 July 1996 22:48, Ben Holiday[SMTP:ncognito@gate.net] wrote:

> I've a few machines around that could be dedicated almost full time to the
> task. What are the bandwidth requirements? Specifically, could the
> keycracker be run over a 28.8 (with a 486 running linux)?  If so, how many
> 486's could I get over a single 28.8 (i.e. 28.8 -> multiple 486's daisy
> chained with ppp over direct serial connection)?

It's not a factor of the bandwidth, you search offline and send in your
results to a central server.

But first, a little reality check is in order.  According to libdes,
the 200Mhz Pentium Pro on my desk will do 1,827,997 ECB bytes/sec, or
228,499 ECB blocks.  A DES crack would have to try, on average, 2^55
blocks.  That would take my machine 43,798,875 hours, or 1,824,953 days.

OK, so let's be reasonable and say that a week would be a good time to
come up with a DES key.  We would need 260,707 200Mhz Pentium Pro's to
achieve this.

Looking at that, 30 days seems not such an unreasonable target.  We would
need 60,831 200Mhz Pentium Pro's to achieve this.

It seems obvious to me that DES is still *way* out of reach of anything
other than special purpose hardware.


Regards,

- Andy (hoping he got his sums right)


PS. For those more acquainted with Sun hardware, an Ultra-1 will do
    1,683,647 ECB bytes/sec (gcc 2.7.2).