From: koontz@MasPar.COM (David G. Koontz)
To: cypherpunks@toad.com
Message Hash: 02e6f22d9a8f1634e7eba238b80637cdd00acb8341bdd4280c1887bd721ef3a9
Message ID: <9508181832.AA21899@argosy.MasPar.COM>
Reply To: N/A
UTC Datetime: 1995-08-18 18:28:48 UTC
Raw Date: Fri, 18 Aug 95 11:28:48 PDT
From: koontz@MasPar.COM (David G. Koontz)
Date: Fri, 18 Aug 95 11:28:48 PDT
To: cypherpunks@toad.com
Subject: Re: Netscape security
Message-ID: <9508181832.AA21899@argosy.MasPar.COM>
MIME-Version: 1.0
Content-Type: text/plain
>Now, I can see calling a MasPar a "parallel supercomputer"; another effort
>at the SSL challenge got the answer about 2 hours before Damien's did,
>and used about 4 days of spare time on the MasPar. Last time I looked,
>a MasPar was selling for about $150K, though I don't know how big the one
>used on SSL was. At that price, you could have your own for ~$500/day,
>and ripping off $2000 on a credit card isn't tough in today's automated world.
>Next year - computer time costs half as much.
Well, lets put it this way, an MP-2 with 16K processors could attack a 56 bit
key about as fast as a 486 could attack a 40 bit key. The next generation
MP will be less than an order of magnitude faster, although able to use
more processors.
An MP2 with 1K processors can do 220K crypt(3)/sec. Now, if it weren't so
darned hard to program one for performance...
One tends to wonder if there are analytical methods to reduce the search
space.
Return to August 1995
Return to “koontz@MasPar.COM (David G. Koontz)”