[rc5] keyspace flaw

Tristan Horn tristan at ethereal.net
Thu Jun 5 14:09:21 EDT 1997


On Thu, 5 Jun 1997, Benedikt Eric Heinen wrote:
> > Now, can we please kill this thread?

I was referring to the conversation about recycling blocks.  There is no
point to this, as far as I can see.  If we were competing, it would only
help the other effort if we ensured that all of the keyspace up to 0x0C
had been searched by us.  (That is, until we collided -- e.g., the other
one was going down from 0xFF and we passed each other in the middle at
0x8.)

> > PS: I also have one good reason why sequential blocks are better than
> >     random ones, so don't even think about bringing that up... :)
> 
> I'm sorry, but there's something going wrong here...
> 
>   At first, I try to give arguments *FOR* random block distribution. Some
> people come along say this is not a good idea, since it would decrease the
> chance of cooperation between bovine and possible other efforts.
>   Following that, I come up with an idea to improve the situation for a
> possible cooperation, and all I hear is 'can we please kill this thread'?
> Is there any special reason *AGAINST* constructive comments? ;)

In this case, I think you mean competition when you say cooperation...

I don't foresee anyone trying to compete with us unless they have a really
good reason (e.g. they have developed RC5 cracking chips which collectively
run 5 or 10 times faster than us) -- otherwise they're just duplicating
effort, IMHO.  If nobody is going to compete with us, it doesn't matter
what parts of the keyspace we search.

Regarding random vs. sequential block distribution:

- statistically, we have a 50% chance of finding the key either way.
- it's already sequential -- why change?  it may require additional
  effort to implement (see below).
- with sequential block distribution, we don't have to worry about
  expiring keys.  someone can grab 3000 blocks and take as long as they
  want to finish them.

I thought I had more/better reasons than that, oh well.  I don't remember
your original reason(s) for suggesting random blocks -- please remind me
if I haven't covered them here.

I'm all for constructive comments, it's just the non-constructive
(destructive?) ones I don't care for.  (now I'm getting a Mac vs. PC war
on two mailing lists.. sigh)

Tris

----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.



More information about the rc5 mailing list