Slashdot Mirror
distributed.net Contest Setback
meisenst writes "distributed.net is reporting that there was a problem earlier on in the CSC project (a few weeks ago) and that about 25% of the blocks will have to be re-done. Their announcement is here." We've gotten more than a few submissions about distributed.net showing more than 100% of all packets processed, but held off posting it until we had the official word. And here it is.
distributed.net is made up of humans, people make mistakes. It's not the end of the world, but it is a little annoying.
On the plus side, with the current key rates the remaining 25% will be cracked within a week or two.
The other plus is that the key may not be at the very end of the remaining 25% of the keyspace and therefore we will finish even quicker!
In Soviet Russia...michael would be rotting in Siberia!
Its only about a week of additional effort... probably the most hit will be the RC-64 project, which has been diminishing its rate slightly recently because of people flocking to CSC... hopefully, just another week is needed before getting all the people working back on RC-64.
Go distributed!
It's VERY important to note that "Official" work about the CSC keyspace being over 100% was ALREADY explained!
It works like this: D.Net rechecks blocks from suspicious clients. D.Net rechecks they by handing out those blocks again. D.Net has said repeatedly that they give full credit for EVERY block handed out in this way from the keyservers. AS SUCH it's easy to have the total go over 100% - this had NOTHING to do with the server problem. The article posted does not make this clear, and should be revised!
Hey Rob, howabout that tarball!
Oops... Another 24 hours now...
"Going to war without France is like going deer hunting without your accordion." - Jed Babbin
I wonder how many barrels of oil that wasted.
-- Virtual Windows Project
There is actually more than one problem with CSC, which is causing it to go over 100%
.plan, stating that 9-12% of the keyspace was being duplicated.
First: dbaker (Daniel Baker), released an official anno uncement explaining that the same blocks were being issued to multiple clients, to attempt to detect cheaters.
Then dbaker released another anno uncement in his
Second: nugget (David McNett), released an announcement stating that there had been a problem with the keymaster generating invalid blocks, resulting in 25% of the keyspace being duplicated.
So, one remaining question is, are they still sending out ~10% 'verification blocks'? Or have they abandoned that to allow us to complete the project faster?
We have reached 112% due to verification blocks and could reach 140% due to 25% of the keyspace being corrupt. However, if 12% of the 25% new blocks are duplicated, then we could reach about 155%...
--
David Taylor
davidt-sd@xfiles.nildram.spam.co.uk
[To e-mail me: s/\.spam//]
The invalid keyblocks have already begun redistribution, so yes they will be reprocessed (correctly this time). The person who correctly resubmits a validly re-computed version of that block will also receive credit. It's not possible for the old previously distributed version of the block to capture the second-pass credit for the block because the invalidly computed blocks are now being screened and discarded as they arrive at the keymaster.
Don't waste those cycles! Put them to use! http://www.distributed.net/
While most of dnet's client software is open source, the keymaster and proxy software is not. The problem that was found was found in the closed sourced keymaster software. This bug may have been fixed sooner if the source were available. I don't understand why the source needs to be closed for the keymaster software. Proxy and networking, I understand the arguments for (though I do not necessarily agree with).
TomG
... to nugget. In all honesty, it can take a lot to admit that you fucked up, and that's what this is. As he said; they're human too, and mistakes happen. We should just be glad that this got caught before we were at like 200% and wondering what the hell was going on :)
Now let's get cracking and finish this thing, and my applause again to nugget for getting it all fixed and admitting it.
---
Tim Wilde
Gimme 42 daemons!
If they were to release the keymaster source, or the client source as written somewhere above, there could be rampant abuse of the source in order to produce what some people would think of as desirable results.
We've seen this before, actually, and it's still in the process of being dealt with; the release of the Quake source to the hungry open source world. Immediately, those few (it's always the few) who have suspicious morals began to use the Quake source to cheat, effectively giving themselves impressive handicaps.
The same thing could most certainly be done with distributed.net source, especially if one knows how to trick the keymaster into thinking certain things (a block is done without really being done, a client submitted 100 packets instead of 10, who knows). It is for this reason that I not only thank distributed.net for -not- releasing their source, but applaud their decision to not do so.
Sure, I'd be more than happy to know how they do what they're doing so well, and yes, the bug would probably have been noticed/squashed more quickly had there been a few thousand code monkeys jabbering away at it, but to be quite honest, the distributed.net team responded as quickly as they could, and in a very professional manner. So, there's really nothing to complain about, IMHO.
meisenst
Green's Law of Debate: Anything is possible if you don't know what you're talking about.
1) If you were trying to cheat, we've noticed you and your efforts are ineffectual.
2) If you somehow have a broken client, read-only disk, or have run out of disk space and are inadvertantly submitting duplicates, you might want to check on it. You might be wasting cpu or network bandwidth that you are not aware of.
Don't waste those cycles! Put them to use! http://www.distributed.net/