Re: DES & RC4-48 Challenges

At 11:52 PM 8/21/95, Timothy C. May wrote:

>Call it a factor of "only" 6000 times harder than the SSL challenge. Hard
>to imagine this happening in the next two years.
>
>Maybe if much of the Net community was energized to run DES crackers
>instead of Flying Toasters, but a hard effort to organize...for fleeting
>reward.

        Given the rate at which news of the prior cracks seems to have
spread among people quite new to these questions, I think you'd be
surprised: I've heard mention of it from no less than ten people who, to my
knowledge, had never before taken any interest whatsoever in crypto
questions. Granted, ten people a-laboring away on Pentiums and PPCs ain't
much--but, who knows?, my experience might just scale quite well.
        Yes, I know: Life is short and art is long. Still, I think it's
worth a try: failure seems likely and success remote, but how much sweeter
victory if the project were to succeed.
        The key, I think, would lie in making participation in the project
extremely accessible: developing simple platform-specific apps that'd make
sweeping space nearly idiot-proof. If joe.anne.net could DL an app
appropriate to hir platform then fill out field in a web page that would
delegate keyspace according to the question "I can let my [platform] run
for [n] hours," and easily report back the results, the response might be
quite strong. How long it would take to succeed, _if_ it did, is anyone's
guess: it could be a day or a decade.
        Obviously, the preparation would be a labor-intensive; the
trade-off, a good one imho, is that this labor having been performed, the
reservoir of potential contributors would expand manifold. If we could
increase the reservoir by a factor of 1000, which isn't at all unlikely,
that advance would be nothing to sneeze at.
        Cracking something that for now seems beyond reach would up the
ante in a pretty big way, and would put that much more pressure on
policymakers to jack that bit-limit up. And that's exactly what we want.

Ted

[stuff about running crypto screensavers]

It seems like there's got to be an easy way to divvy up keyspace on a
real time basis so that anyone with a few cycles to spare can
contribute them...

There could be a central server out there, which would keep track of
what keyspace needs to be swept, etc.

Then, clients (available as Mac, DOS, Windows*, UNIX,
etc. screensavers) could, whenever possible, connect to the server,
get a bit of keyspace, and start checking.

The client would tell the server how fast it is (chip and speed), and,
optionally, an estimation of how much time it will have free.  The
server finds some unchecked keyspace and allocates the range.  Then,
when the screensaver ends (when the user starts working again), it
would check in and tell the server how much of the space it checked.

The server could implement multiple passes of the keyspace to cope
with untrusted clients, etc.

Or something like that, anyhow.

The point is, the keyspace should be divided on a spare-cycle basis,
so that people who happen to have some extra, unused time on a
workstation can help without having to follow this list religiously,
and without having to commit to checking a certain amount of keyspace
beforehand.

This really is a nifty concept, and I think a *lot* of people would be
interested in running crypto-cracking screensavers.

--sq

On Tue Aug 22 12:26:49 1995: you scribbled...
> 
> 
> [stuff about running crypto screensavers]
> 
> It seems like there's got to be an easy way to divvy up keyspace on a
> real time basis so that anyone with a few cycles to spare can
> contribute them...
> 
> There could be a central server out there, which would keep track of
> what keyspace needs to be swept, etc.

It seems that the SKSP protocol is a good foundation for just such a plan.

Can screensavers such as After Dark for mac/pc access network protocols?  
or would someone have to write a new screensaver alltogether.

...alex...

Alex Tang wrote:
> 
> On Tue Aug 22 12:26:49 1995: you scribbled...
> > 
> > 
> > [stuff about running crypto screensavers]
> > 
> > It seems like there's got to be an easy way to divvy up keyspace on a
> > real time basis so that anyone with a few cycles to spare can
> > contribute them...
> > 
> > There could be a central server out there, which would keep track of
> > what keyspace needs to be swept, etc.
> 
> It seems that the SKSP protocol is a good foundation for just such a plan.

 > Can screensavers such as After Dark for mac/pc access network protocols?  
> or would someone have to write a new screensaver alltogether.

Afterdark is just a program that always runs in the background. You
could program a screensaver/timeslice stealer...even by modem...

And with Windows new 'features' you can get someone to install this
at the click of a button..from an email message, Ha!....