Slashdot Mirror
Distributed.net CSC Success
dbaker let us know that distributed.net has finished another one, half an hour ago as I write this. IIRC, they had to redo part of the keyspace, right? Looks like it didn't slow the project down too much.
← Back to Stories (view on slashdot.org)
I'm really glad CSC is done, because the Mac OS CSC client sucks. Doing CSC, my P/233 running Linux 2.2.14 was outperforming a G3/300 running Mac OS 9. That was just scary. Plus, on slower Macs, the CSC client lagged the whole system really badly.
But that's all done now. The rc5 client is fine. It even supports AltiVec - makes me wish I had a G4!
Hmm, I guess this is kind of off-topic rambling. Oops. I'll stop now.
$x='S24;r)>63/* h@<5+oZ)32"5cz';$me='phroggy'x$];
$x=~y+ -xz+\0-Tx+;print$_^chop$me for split'',$x;
I'm finding encryption challenges progressively lamer. What are we supposed do be doing? Besides trying to win money with chances which rival that of your average US state lottery, I think we're trying to prove that the encryption that the government wants to restrict us to is weak. The DES projects have proved this - for single pass 56-bit DES keys, the code can be broken quickly if you have a huge amount of cracking power available to you. The CSC project has proved that it's a little harder right now but it can be done. The RC5 project (which is pretty useless at this point), has already proved that if you're using RC5 encryption with a 64-bit key length, for the moment, you're safe unless you're paranoid about strange people with phenomenal amounts of cracking power. Is this what we want to prove? that 64-bit keys are good enough? I was under the impression we wanted free access to 128-512-bit keys for everyone.
The one thing which still interests me is the use of these contests as a way to chart CPU power over time. The RC5 project is currently about 17.5% complete, after 815 days - but look at how much faster the second 1% went compared to the first 1%. This is (among other things) Moore's law in action. The one thing that disappoints me about the CSC project is that there never got to be a AltiVec (Velocity Engine) enabled version for powerpc G4s. The RC5 client now has a AltiVec version and it's about a 4x speed improvement. That's impressive.
I think projects like GIMPS and OGR are far more useful and worthwhile. I don't like SETI because of the especially closed method of the project and the rather arrogant idea that looking for something, which may or may not be there, in a way that may or may not find it, while we may or may not understand what we're doing, is useful.
I will continue to crack RC5 blocks until something better comes along.
I'm not a smorgasbord.
Actually they have a damned good idea what it will look like, a narrow peak. These are not known to occur in nature, but any technology using the radio spectrum for communication will cause these.
That's only true using the most simple transmission methods. Enter spread-spectrum or any of the carrierless transmissions and you lose that peak.
Incidentally, there are also projects searching for those very laser-light signals you mention, as well.
Um... wouldn't a laser signal have to be pointed directly at the intended target? Yes I know all about divergence and all those great things but looking for laser communications perhaps not intended for us seems rather silly.
Most excellent!
Congrats to the guys at Distributed.net, the guy who submitted the winning key, and his team (if he was on a team).
Now we can get back to that darn RC5-64 contest.
I really wanna see what some of these new systems hitting the market will do (1Ghz Kryotech Athlon system?).
Anyone else built a big-arsed cluster to work on this yet?
Somebody sneak a client on a disk into the Fermi-Lab cluster! =)
Chas - The one, the only.
THANK GOD!!!
Chas - The one, the only.
THANK GOD!!!
SETI@home v2 should be released soon. I hope anyone who feels as you do would give it another chance - they've redone a lot of useful stuff in the software...
My SETI(@home) pages and club info.
I have about 55 computers in my sphere of influence. 20-something are running D.net's new client. About 5 are running S@h. The others are portables and are running nothing, or a CPU cooler. There's room for everything...
Check in a week or so.
This is serious. I mean, if I thought for one second that all we (and the universe) had to look forward to for the rest of eternity was the human race spreading like a virus, I'd kill myself. The hope that we'll meet something fundamentally different from ourselves, some time, keeps me going. (The hope that it will be in my lifetime, keeps me productive ;)
I personally think that Cosm is the "Right Way" to do distributed computing, but I'm not a programmer, so I can't help them code it. However, if they actually get a client up - I'd run it, and I'd promote it as the right way to manage IDLE-time projects. I'm sure that heaps of people behind obscure corporate firewalls would be ecstatic to have one program that manages they internet connection and data transfer...
Hey, I wouldn't mind getting my name on the list of people who've discovered Mersenne primes, especially since it includes people like Euler.
I do it because it's fun, but there's also a 10,000 Euro prize for the winner. Of course, some of this money goes to distributed.net to maintain the network, some to charity, and the remaining to the winner. Still, 2,000 Euros is nothing to scoff at.
There are other project coming, such as the OGR project. It has nothing to do with encryption.
Phobos - Greek word for fear or flight
Finding primes: The only reasons are fame and money. Do they find anything? Yes, but at essentially random times.
:-).
Finding keys: The only reasons are fame and money. Do they find anything? Yes, at highly predictable times (hence making the actual finding a trivial matter).
Finding extra-terrestrial life: Fame, money(?), one of the greatest scientific and philisophical questions in history answered. Do they find anything? Nope, and they aren't likely to anytime soon for various reasons.
So you can either have useless+sucess or useful+failure.
Someone mentioned a future distributed.net project that was not key cracking and actually had reason for existance. When i am not so tired, I will look into this, but it really looks like the way to go.
Incedentally, I run two copies of Prime95 and the Seti screensaver on a 4 hour delay. Draw conclusions about my personality from that
Why do you assume that distributed.net will never do anything but cracking encryption? They're working on setting up a way to distribute the calculation of Optimal Golomb Rulers, a mathematical concept that has proven real-life uses.
And what good would it do to give more CPU resources to SETI@Home at this point? Triple- and quadruple-checking their limited data doesn't seem very efficient to me.
A distributed calculation of pi would be cool, but are there any algorithms to find digits of pi without having calculated the previous digits? If not, it can't be distributed.
--
Win dain a lotica, en vai tu ri silota
CSS keys can be recovered in seconds (and have.) See the LiVid mailing list archives (www.linuxvideo.org) for several explanations, example code, and a large block of keys recovered with the code.
CSS is laughablly weak.
The issue at hand is not that CSS is "simple" to break. The issue is, should a CSS case be brought up in court by the Big Bad Lawyers, what would happen?
Actually, I think the main issue is that by breaking CSS this way we can redistribute DeCSS with all the keys without facing a reverse engineering charge.
Jay
_________________________
Believe it or not, there are some people who don't believe in extra terrestrial intelligence. And those people aren't all psycho fundies either. We choose to work on Distributed.net because to us, it's far more purposeful than seti at home, which would be like looking in an empty box for...something.
We could also run Prime95 but I would rather discover how easy (or hard) it is to crack encryption than waste my cpu cycles looking for a bigger prime number.
So, given the choice between looking for something in an empty box, or finding the next big prime number (which does, um, nothing for number theory) we choose to crack codes, which might prove useful (primarily in discovering what kind of computing power is necessary to do it).
Jeremy
Looking for a Python IRC bot?
But what if radio signals are bad for their genetic makeup, which is why they don't use them and aren't answering our calls? Maybe Seti is killing them slowly! And all of us using the client are accomplices to murder. Maybe they're on their way here right now to kick our sorry little Seti-asses. Or maybe they're trying to mentally weaken us through TV(our great weakness.) Good example:Professional Wrestling. Or maybe radio waves are good for there libido, and they're coming here to thank us for helping to perpetuate their race. Or maybe its late and I need sleep.
Yeah, last I heard they were doing that until they can get more telescope time. They've also released newer versions of the client which do some more checking on older data.
Adam Beberg was one of the founders of distributed.net. While their first contests were all encryption related Adam really wanted to pursue general distributed computing problems. v3 of the D.net client was supposed to have a plugin type model which would allow people to run one client and pick the problems, be it encryption, primes, ogr or even seti. It seems that the timing was off because Adam decided to leave d.net to work on cosmo, the client that would have been v3.
Why can't everyone stop fighting over which project to support and show enthusiasm for cosmo. Once it has wide spread use much more work could be completed. Each new problem would not require a whole new client. People could easily work on SETI when they have unchecked data and look for large primes, or fit circles into boxes the rest of the time. Currentlt each project requires a seperate download.
I would hope that CPU resources that currently valuable or needed are not being used on distributed computing projects like these. That thought aside, pi most certainly does NOT have a repeating pattern. It is irrational (cannot be represented as a fraction of two integers) and transcendental (is not the root of any algebraic equation). The proof of the transcendence of pi is involved, but the proof of its irrationality is significantly simplier, relying on basic integration, infinite series, and one trivial property of asymtopic growth of recurrence relations.
Try http://www.seanet.com/~ksbrown/kmath313.htm
Cracked Code wins challenge
Winner did use Solaris
and I think, 'So what?'
They've also released newer versions of the client which do some more checking on older data.
Um, no they haven't released newer versions. According to press releases they were going to release the new version (2.0) sometime last week, but it never showed up and nobody on the project ever said anything about it. The fundemental lack of respect or responses (which this is only one example of) from the team running seti@home is the reason I don't donate my time to it anymore.
Perhaps its time to reallocate our valuable and abundant CPU resources to things like SETI@Home or perhaps a project to distribute the calculation of pi to see if we can find a repeating pattern. Either way, distributed is an obsolete and top-heavy project that has served its purpose. Maybe its time to put our hacking minds together and come up with something altogether new to think of. Perhaps theoretical artificial intelligence.
; }return(0);}
Lets reallocate all this power to productive uses.
#include <signal.h> \ #include <stdlib.h> \ int main(void){signal(ABRT,SIGIGN);while(1){abort(-1)
OFTC: By the community, for the community
I think the other posters to this comment are misunderstanding the author's point.
The issue at hand is not that CSS is "simple" to break. The issue is, should a CSS case be brought up in court by the Big Bad Lawyers, what would happen?
It is understood be the technical community that CSS encryption is pretty much useless as a copy protection scheme. But how can this be proven to a jury? Expecially when $500-$1000 per hour lawyers are employed by the adversary?
distributed.net is a known and proven source for ranking the status of encryption. They are not a malicous orgainization, but rather a testbed, placing an encryption technology under intense scrutiny and forcing it to prove it's worth. Where CSS (or its relacement) placed under such scrutiny, the results may be enough to have future lawsuits thrown out of court. (and therefore lessening the slashdot posts/comments on this subject... no offense, but it is getting kind of old. There is very little new information on the subject left)
"You want to kiss the sky? Better learn how to kneel." - U2
Sig:
Barbeque is a noun. Not a verb.
Expanding on the distributed.net idea, would it be practicle to create a more general client for distributed computing?
Here's what I mean:
The client would be written to actually accept algorithms via the internet and then accept the data to be crunched.
Why?:
Lots of people working on lots of computers and lots of power intense applications (I'm thinking along the lines of academic research, but this could be applied within corporations, too).
Premis: Those CPUs, some big some small, are not all busy all the time. With this client out there, those spare cycles could be used by others within the CPU-sharing-organization.
Example: All the computers in the Math department. Or, on a larger scale, several universities could coop thier super computers.
Sorry, I've just been drinking too much coffee, too late in the day...
Don't you think it's time to start communicating?
I thought it was talking about distributed.net cracking CSS code, instead of CSC.
:)
For the on-topic part:
I think it's good that distributed.net's accomplished another goal. Just showing people what the community is capable of is an end unto itself. An earlier poster commented that they believed it was a waste of CPU cycles, but the whole idea behind such an effort, if I'm remembering correctly, was that those CPU cycles would have been unused anyway. Might as well show off with them
For my idea, which the on-topic stuff spawned:
I'm not a lawyer, so I don't know about the legality of this. What if distributed.net were to host a challenge where people could use their cycles to crack the other keys on CSS, for DVD-playing? Granted, it wouldn't be a very long challenge, as CSS is fairly simple, but it would be a nice, big statement about how the community in general feels toward the recent actions of the MPAA and DVD consortium. Probably not legal at all. Still, it would be a fun thing to do.
-Denor
Useless? What is your CPU going to be doing otherwise? I've found that it hardly interferes with anything on my NT box, and the Linux server which HAS to be up and operational for those occasional logins is not doing anything besides using up precious Kinetic energy. Might as well be doing something.
Linux . Because a computer is a terrible thing to waste.
Distributed.net - Because CPU time IS BEING wasted.
It might look like it didn't slow the project down too much, but a clear-headed person would not be fooled.
Oh, look, I just flipped a coin and it came out heads. Obviously, this coin doesn't come out tails too much, does it! Wanna bet me?
That's why I like to run SETI@Home. With an encryption challenge, you know it's gonna break eventually, it's just a matter of permutation. SETI is much more exciting (and worthwhile) IMHO. For those of you who don't know, SETI@Home uses the Arecibo radio telescope to look for extraterrestrials. Work units are sent to people running the client which checks for things like gaussian curve spikes and other abnormalities. Also, they have clients for just about every platform you can think of.
Exciting: most people who compute work units for speed turn off the friendly screensaver, so I don't see how it's more exciting.
Just kidding; I understand what you mean, but I do disagree with you. SETI is a program with a noble objective, however, it is fundementally flawed, both in implementation and premise.
Implementation: checking work units multiple times (ie. many more than would be required to detect cheater clients) because of distribution/data collection inefficiencies does not neccesary seem very productive to me
Premise: Yes, it is almost certainly true that there are extraterrestial beings somewhere "out there." Nonetheless, there are two main obstacles that make it exceedingly unlikely that we will ever find anything: space and time. Space is pretty obvious; it's a big universe out there and radio signals, if any, will take a long time to propagate. Time is the big one: the probability that another civilization is at the same technological stage as us such that they use radio to communicate in astoundingly low. Just imagine: 200 hundred years ago, a mere drop in the bucket in geologic time, we could not even receive radio signals. 200 hundred years from now, whose to say that we won't be using fiber optics and laser relays to transmit information. Two cultures that are at different levels of technological accomplishment can rarely communicate or interact: look at the much smaller separation in time between the Europeans and Native Americans and the disconnect that occur then.
Seriously, I hope that they do find something. It would represent a milestone in human history that will force us to reconsider our perspective in the vast scheme of things. But I just don't think it's going to happen.
Maybe I've been hitting the Viridian's Site a bit too hard, but I believe the biggest problem with Distributed Computing as we currently know it is the amount of energy being consumed by people running the clients. I think that d.net, GIMP, and Seti@home all have things going for them, but none of these is as important as the air that we're breathing right now. I'd rather let my box go into standby, or *gasp* turn it off (when possible) and save a little energy than accomplish any of these goals.
Feminism is the wild notion that women are human beings.
I'd like to again thank everyone for their support throughout this project. distributed.net is nothing without the amazing user support that we receive.
For the record, we received official confirmation and congratulations from CS Communications & Systems on our CS-Cipher solution.
For those that missed it:
Contest: CS-CIPHER
Solution: 61 A3 9E 36 BF 4F F0
The secret message is: CS-Cipher a ete presente en mars 97 a Fast Software Encryption (PARIS). Congratulations to the winner!
Moo!
-dbaker
--
Daniel Baker - dbaker@cuckoo.com - dbaker@distributed.net