[HARDWARE] Just Curious......]]

Dan Oetting oetting at gldmutt.cr.usgs.gov
Mon Sep 20 10:40:25 EDT 1999


At 5:02 +0000 9/20/1999, taniwha wrote:
>harder will be finding test vectors .... you gotta get at least one of those
>false positive  keys that dnet has probably already discovered to check out
>your logic, the last thing you want to do is to miss the key and have all
>of dnet have to start from scratch

If you are going to hard code the chip for one problem it would be just as
hard to find a test vector as finding the real key. If you leave some
parameters as inputs such as the clear text and cipher text you can create
as many test vectors as you want.

Another option is to enter a mask that defines a subset of the cipher bits
that must match after encryption. A test key used to generate the mask can
be secretly picked by a server and the correct result of processing a block
will be to find a key whose encrypting result also matches the mask and is
at least as good as the test key. Finding the real key is no longer a
special case that will only happen once but just a very good match that the
client treats like any other result.

>you need to know how fast it's going to go and you have to get
>dnet to buy in to letting you hook up to the client (from what
>I've heard they are understandably paranoid about clients -
>because they've had a few people hacking them to
>get fake keyrates - and for the reason mentioned above)

It's so simple to hack the current clients I'm surprised the over all
keyrate is still so low. I've presented a plan to insure the integrity of
the RC5 cracking process but d.net has never shown an interest. If you are
going to build dedicated cracking hardware you might want to bypass d.net
and setup your own key server. Then you would be able to control your own
process and won't be worrying about d.net leaving the barn door open.

--
Dan Oetting <dan_oetting at comug.com>



--
To unsubscribe, send 'unsubscribe hardware' to majordomo at lists.distributed.net



More information about the Hardware mailing list