[Cryptography] GPU Farm (Marco Streng)

[Peter Trei]

On Mon, 18 Aug 2014 18:26:49 +0200
Marco Streng <marco.streng at genesis-mining.com> wrote:

>
>
> Hello Guys from metzdowd,
>
> I really like your mailing list and following it longer time now. Since
> I think the posts there are really of high quality and the people
> participating really know what they are talking about I would love to
> ask a specific question that could be really interesting for everyone
> there.
>
> I come from a company that is doing cryptocurrency mining in large scale
> and currently we are thinking about a use case for our 1300 high end GPU
> farm that is not necessarily profit oriented, but most importantly
> interesting and exciting. I asked a lot of people till now and also
> people from chaos computer club in germany, where I received especially
> ideas pointing to using the farm to decrypt emails or other decryption
> services. Also I was suggested to use it to decrypt the NSA_KEY, but
> that is obviously not doable in human time even if the farm would be
> 1000x bigger.
> Do you have any interesting idea what to do with the farm? If we like
> the project it we wouldn't hesitate to start directly!
>
> I would really appreciate if you could ask this to the mailing list and
> I am really excited about the answers. What do you think?
>

Marco:

There's an obvious classic target: The RSA Challenge Numbers.

Way back when the RSA algorithm was just getting established as a usable,
there
was considerable doubt over its security. In 1991 RSA (the company) posted
a set
of  'Integer Factoring Challenges', with cash prizes attached. See:

http://en.wikipedia.org/wiki/RSA_Factoring_Challenge

In 1996 RSA also posted a set of 'Symmetric Key Challenges', with messages
encrypted using DES and various strengths of RC5. I was the proposer of
these challenges, and the brute forcing of DES and increasing key
lengths of RC5 contributed to the relaxation of crypto export rules in 2000.
The longest key brute forced was 64 bits. distributed.net is still working
on
the 72 bit challenge.

The prizes were withdrawn in 2007, on the grounds that they'd served
their purpose - RSA was still 'not broken' at adequate modulus length;
(no general break had been found), and exportable short symmetric keys
were no longer required. (There was another problem that individuals
were getting in trouble for using slack cpu on machines they did not own
(ie, their employer's) for brute force calculations).

I don't see the Symmetric key challenge messages up on the RSA
site, but I could probably track them down.

Forcing a symmetric key isn't interesting at the moment - most
responsible sites are using at least 128 bit keys. However, 1024
bit RSA keys are still in wide use, and factoring the RSA-1024
challenge number would pressure people to update their systems.

Sadly, it won't get you the $100k prize, but the publicity would
be valuable.

Peter Trei
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20140819/8990f085/attachment.html>