Subject: Re: 1024 bit key considered insecure (sshd)
To: None <tech-security@netbsd.org>
From: Seth Kurtzberg <seth@cql.com>
List: tech-security
Date: 08/29/2002 19:02:42
The fact that a massively parallel attack can succeed is hardly new 
information, as Perry noted.

For a little perspective, I'm installing a NetBSD based firewall this weekend 
for a company that has credit card information in their database and had no 
passwords.  Not no encryption, no passwords!  With juicy targets like that, 
is anyone really going to be spending large sums of money to break into my 
data?  Actually, I hope they do because I keep a windows 2000 server running 
IIS that is full of phony credit card numbers.

On Thursday 29 August 2002 12:27, you wrote:
> "Dave Feustel" <dfeustel@mindspring.com> writes:
> > See _Cracking DES_
>
> Wow. I'd never heard of that book before. I wonder why no one
> mentioned brute force attacks on DES to me. It might have been
> interesting to mention to my students in my annual graduate course in
> cryptography.
>
> > 
> > (http://www.amazon.com/exec/obidos/tg/detail/-/1565925203/qid=1030639763/
> >sr=1-1/ref=sr_1_1/102-5391104-6813765?v=glance&s=books
> >
> > for a (by now obsolete) low-cost home-brew system
> > for cracking DES. The available FPGA hardware has advanced
> > considerably since this book was written.
>
> Don't try teaching grandpa to suck eggs.
>
> For extra credit, present the difference in computational complexity
> between cracking a 56 bit DES key and factoring a 1024 bit
> integer. And no, the difference is not a factor of 2^968. You should
> especially go to the back of the room if you thought it was a straight
> factor of 968, and if you thought it was a factor of 18 because 1024
> is about 18 times larger than 56 you should confine your future job
> searches to the food service and waste disposal industries.
>
> For extra extra credit, figure out how many Virtex II FPGAs you would
> need to try out Dan's new number field sieve trick with a 1024 bit key
> if you want a result in one year. The Virtex II is ideal because of
> its size and the presence of several IBM PPC cores on board. Hint: it
> is not clear Xilinx can produce that many Virtex IIs for you at the
> moment, though I'm sure they could scale up production for it.

-- 
-----------------------------------
Seth Kurtzberg
M. I. S. Corp.
1-480-661-1849