Slashdot Mirror
Seti@Home Now Has Teams
Madoc writes "Was just over at Seti@Home's site, and saw that they've introduced teams now! There are 2 Slashdot teams, we should probably standardize on one:
Slashdot.org
and
Team Slashdot "
I vote for Team Slashdot. Go seek out new intelligence if
this rocks your boat better than cracking DES keys.
Yes, it would be nice to finish one contest before moving on to the next, but look at the RC5-64 stats! It's been running for 577 days and we're only through 8.5% of the keyspace. Now I know, we're cracking at speeds faster that ever before, but still, it's proven to be fairly strong crypto and all we're doing now is wasting distributed CPU power which can be used for much more usefull projects. I've personally switched over to SETI@Home ans I've noticed a lot more press about it than the RC5 contest. The reason? It impacts a lot more people than encryption. How many people care about strong encryption. Ok, now how many people have stared up at the stars wondering if we're alone in the universe?
One of my (two) biggest complaints with the seti@home project thus far is that most aspects of it have been under-engineered. Not enough time has been put into developing any aspect of it, save possibly for the cruncher algorithms themselves. Distributed.net did a far better job of delivering clients for platforms, assembling a good server-side package, proxying of many sorts, copious configurability, SMP and CPU-specific processing cores, etc.
All that said, the SETI project does stand to yield something more useful, at least psychologically -- the time spent beating on RC5 has mainly (and successfully) demonstrated that DES-56 sucks, and that bigger keys are vastly harder to break. If there came another rapid DES-breaking project such as DES-II or DES-III, I'd happily switch my spare CPU cycles back to it for a day or two.
Also, the source to seti@home isn't available, a problem which they have yet to rectify. If they desperately need to protect the algorithm for scientific integrity, they can move all that to a library and open the rest of the source so that we can fix the missing parts.
Wow, just imagine the implications... Joe Celery's computer receives the block, the one block that has a transmission from intelligent life forms. It's a hot day outside, and his "kewl 504A" is just a tiny bit hotter than when he left it running the Unreal timedemo all night to see if the chip was stable. Due to a random heat-related glitch, his computer mistakenly reports the block as not containing anything interesting, so it goes unchecked.
Five years later, an alien demolition team wipes out the entire Milky Way to make room for an interstellar frontage road, a procedure that they had advertised (via radio beacon) for millenia. All life on Earth perishes because of an overclocked 300A, whereas if Joe Celery had not overclocked his chip, humanity would have made first contact with an alien race and Joe Celery's name would go down in the history books.
Now wouldn't that just suck?
Wanna know how do do this? It's so easy it's insane.
Let's say you want to add Michael Dolan (top individual user by miles) to your team. You get his email address (which is clearly displayed in the top 100 individual list) and change the logon in your Winbloze client to this email address. The user info in the client display will now tell you that you are Michael Dolan.
Now look in the program directory for "user_data.txt" or some such. Open it up and look for the "key" value. This is Michael Dolan's password! Bingo, you now have his email and his "password" - add him to any team you like, or all of them if you want! Wanna have more fun, add the top Big companies to a team, or Berkeley themselves! Ha ha!
Distributed password in plain text format? How STUPID is that?
Two points:
1. The statistics for CPU time for Windows 9x include the time it is staying minimized in a System Tray doing nothing.
2. To speed up processing time in Windows 9x dramatically (about 3x in my experience), turn on screen blanking in the screen saver properties.
Account maintenance should be finished soon.
-Peter of the SETI@home team
Run it in screensaver mode and configure it to blank the screen. It goes 3 times faster when it's not bothering to draw the pretty pictures.
- A $60,000 machine built by the EFF beat out all
the King's horses and all the King's men
(otherwise known as distributed.net).
- The NSA probably would have considered Deep
Crack (the EFF's key buster) a keen and useful
computer -- twenty years ago.
So, unless you've got some really serious reason to think otherwise, I'd stop worrying about a few bits from SETI, take my medication and start looking for little green men like a good little member of the Collective. Besides, there are better things out there to worry about, like the war in Kosovo or a 1 cent increase in the price of a stamp.
----
Every year during my review, I just pray the words "slashdot.org" aren't mentioned.
My common sense tells me that if someone wants to send back falsified results, they'll send back falsified results. It's been shown time and time again that OBSCURITY != SECURITY. Just because the SETI project is closed doesn't mean that their results are not being falsified at this very moment, because it just takes a little more dedication to screw it up. The only really secure protocol/program/anything-else is one that's been peer-reviewed and shown that it's secure, which means things like checksums, encryption and accountability. It's my opinion that the SETI program is in fact more vulnerable to cracker efforts because it's closed -- a vulnerability in the system, once found, will probably not be brought to light before the results are completely and horribly skewed - after all, if they don't give us the code to review, why tell them when there's a problem in it?
Posted by Kallahar2:
search for XSETI, it's a GTK GUI for Seti@Home
Sure, but I think that the rc5 project has already proven it's point. Even if we do finally finish rc64, all we will have proven is that it is crackable, but that it takes even distributed computing a really long time with today's technology, but I think that this point has already been adequately proven already by distributed.net. If they do finally crack the rc64 challenge, I don't really think it will add to that point at all. I think that it is already understood now that it can be cracked, and that the average time it would take to do so is really just a fairly simple mathematical exercise complicated only by unknowns like the increase in processing power from year to year, and how many people participate as time goes on. Actually doing it at this point holds little more point than the prize money in my opinion. Don't get me wrong, I think this was a vital project when it began, I just think that it accomplished what it needed to do, that's all. Now if they ever get their OGR stuff going, I might be tempted to switch back to work on that for a while, but for now I'm sticking with SETI.
Shawn Asmussen
Let me guess, the plans for the frontage road were on display at the local planning department in Alpha Centauri for fifty years. Or were they actually in the cellar of a planning ofice, at the bottom of a locked filing cabinet stuck in a disused lavatory with a sign on the door saying "Beware of the Leopard"
--- Linux... a college project gone horribly right
I'd rather see more intelligent life on this planet before I spend my energy looking for it elsewhere. We have some short sighted morons making emergency presidential orders about privacy and encryption may be used to aid pedophiles and terrorists? There are a few serious flaws in that logic and reality and its a matter of principle. Until we can enlighten or remove the figureheads in office, I will waste my 60 watts of idle processing power to make a sad political statement.
I would rather look for intelligent life elsewhere, but I think it is more urgent to look for it here first.
Suppose you're a government agency, and you get hold of some important encrypted data. No problem -- just dump the key into the seti@home processing queue. Instant free cycles from enthusiastic geeks all over the country, of whom many are privacy advocates who've been participating in various distributed cracking challenges over the years in attempts to protest your authoritarian policies. O, sweet irony.
Dan Wineman
Which one of the two systems is more useful? Distributed.net or Seti? I think Seti wins between those two... Though, isnt there another one of these types of things that looks for extremely large prime numbers? That one is probably the most useful out of them all.
We've also got a Team on SETI@home. You can find out info about it, along with tips on optimising your SETI@home client software on the Club Team homepage.
Enjoy,
Kris.
Win a Rio (or join the SETI Club via same link)
I have an AMD K6 233 Mhz, a P166, and a 486. I've been running rc5des on my AMD K6 and was running SETI@Home on the P166 and 486. I noticed the P166 and 486 seemed really lagged when connecting to them (even though they where connected via 10BaseT Ethernet). When doing a top, SETI@Home was using majority of CPU (how suprising), but a LOT of RAM too. This was on both machines, which really seemed to suprise me. On my AMD K6, rc5des was using almost no RAM.
Has anyone else noticed this? I'd like to know to see if it's just me or not. Because if it's not, I can wait for new clients and hope that the RAM usage is less.
ZZWeb.net Web Hosting - http://www.zzweb.net
ZZWeb.com Internet Consulting - http://www.zzweb.com
I don't think you're comparing apples and oranges. The SETI project has the possibility of finding something truly important. The best that distributed.net can do is crack an encrypted test message.
That distributed.net will eventually crack the code is a given. All they'll have proven is that it takes a long time to crack RC5, even with lots of computers. There was never any question that it was possible -- just how long it would take. And now we know that it takes a very long time.
SETI, on the other hand, could discover alien intelligence.
Sure, encryption is a very relevant topic. But is distributed.net?
peter
But the nice thing about Linux is that you can bolt the program down so tightly (separate user, chroot) so that it cannot do any damage - it'll never find my pornography or any of my other dirty secrets ;-) (hmm, me reaches for the man chroot command anyway)
Having said that I think it's not really feasible for these guys to give out the source code, because it allows malicious people to write something that'll send fake packets back saying "okay - I've found nothing". This would be a grossly irresponsible thing to do but I wouldn't rule out a cheat who would want to bump up the team's "block count" up a little or religious fanatics whose beliefs depend on there being nothing out there. Security through obscurity, perhaps, but I can't think of any other way of protecting against cheats.*
Despite that I'm still a little irked off about it myself as I'm forced to sit behind a non-transparent proxy and twiddle my thumbs with a cluster of about ~16 decent machines that are just itching to join in the search for extraterrestrial life. If only I had the source I could have written that proxy bit myself already!
*By the way it's probably only a matter of time before someone actually reverse engineers the program. Security through obscurity has always ended in tears.
Hold your horses, the database server is off-line and has been for many hours. Before there it had been acting funny for many more hours (not recognizing user names, so impossible to set up teams).
Try again tomorrow.
TA
Sure you can write your own proxy even though you don't have the source. /etc/hosts file (and if you have a /etc/nsswitch.conf then set it to check local files before DNS of course).
A netstat shows that the client connects to sagan.ssl.berkeley.edu, and a 'strings' on the binary shows 'shserver.ssl.berkeley.edu' which turns out to be the same as 'sagan' right now.
So just make something that can take the connects from setiathome on port 80, and forward it to shserver.ssl.berkeley.edu port 80 (and the other way). Put this 'something' (which also understands your local proxy system of course) and put it on a computer that looks like 'shserver.ssl.berkeley.edu' for the client, you can do that just by putting a fake entry in the
You can probably do it in Perl.
TA
I second that.
;^)
Team Slashdot, we find aliens and crash wussy webservers.
-- 100% MS-Free as of 4-4-1999, 11:47:38 PST. "The lapdance is always better when the stripper is cryin'" Free Kevin,
Hmmm, make sure you are using the 1.1 client. I was using the old clients from the beta testing in the beginning and then realized that those units don't count. Depending on your processor, you really shouldn't need 50 hours. I'm averaging about 16 hours across the 75 machines I am using.
---
My Pentium-II 300 takes an average of 40 hours of CPU time to process a block running NT. I noticed that i686-pc-linux, the average time is about 11 hours. All of the average linux times are faster.
Is the linux client faster? Or are linux users just running faster computers? Maybe it's all the graphics the Win/Mac versions draw that slow them down.
A couple of the people in the top 100 can do the SETI blocks in 7 minutes. Yikes.
I special hardware involved?
Because the SETI client doesn't multithread in it's current version, that would have to be one honkin' processor. Or a parallelizing compiler.
Sigh. I wish they'd implement a way to edit your setiathome user info. In fact, I think they should have done this before adding the teams - as soon as I give them my email address I'm stuck with my old settings and there's nothing I can do about it.
Gripes aside, I'm still running the client because I think the project is so important.
I've put in about 50 hours of processing time or so under the Linux version so far, receiving 7 blocks up to this point. I have yet to show any results being posted to SETI@Home. I do have an outfile.txt (running around 1.5k so far). After the database came back up, I even tried to get back in using setiathome -login and logging back in. Does the outfile have to reach a 'critical mass' before being sent?
Vrallis