1993-08-18 - Re: World record in password checking

Header Data

From: “Perry E. Metzger” <pmetzger@lehman.com>
To: peter honeyman <honey@citi.umich.edu>
Message Hash: dc5f529f46899c1446c752690dd50f12d1b55df224833ed0be255cc2da026cd6
Message ID: <9308182154.AA21533@snark.lehman.com>
Reply To: <9308182116.AA14986@toad.com>
UTC Datetime: 1993-08-18 21:55:47 UTC
Raw Date: Wed, 18 Aug 93 14:55:47 PDT

Raw message

From: "Perry E. Metzger" <pmetzger@lehman.com>
Date: Wed, 18 Aug 93 14:55:47 PDT
To: peter honeyman <honey@citi.umich.edu>
Subject: Re: World record in password checking
In-Reply-To: <9308182116.AA14986@toad.com>
Message-ID: <9308182154.AA21533@snark.lehman.com>
MIME-Version: 1.0
Content-Type: text/plain



peter honeyman says:
> why doesn't this impress me?  i'll tell you why.  with
> 
> o   a stock version of des (dennis ferguson's), which is written in c,
>     and not optimized for any particular chip or vector hardware
> 
> o   a no-name 50 Mhz 486, which you can buy for under $1,000 at fry's
> 
> o   netbsd, a freely available general purpose operating system
> 
> i have measured 29,000 des crypts per second. 
> 
> now give me a "1,024 node" machine made of of these -- admittedly
> unwieldy, but no doubt a hell of a lot cheaper than a 1,024 node CM/5
> (and a hell of a lot more useful, imho) -- and i can run at three times
> the "world record" rate.

You can do even better if you happen to have 2000 sparcstations which
are idle for 16 hours a day. Myself, I wonder how many machines we
would need for a net parallel DES crack.

Perry





Thread