| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Nutritious Rice for the World
Thread: Slowing down Rice project |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 43
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
How many publicly stored copies of results are thought to be needed to know that it will always be retrievable over the Internet and will not have been subjected to tampering?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It all depends on how many clients are available to store information. It wouldn't be an immediate on-demand type of storage. It would need to be used as a temporary archive for mass storage that could be retrieve when needed over say a weeks time. Think of it almost as a bit torrent. No one computer has the same data stored on it. It would have different pieces but other computers would have the same pieces as well. It would be encrypted so it couldn't be tampered with.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Whole papers have been written on the subject, and there are more complications than perhaps you have considered.
What happens when a host goes offline permanently? You need a mechanism for redistributing the data. Anyway, in most cases, the researchers do need instant access to the data. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
As far as the web service (that I can't remember the name of) it did that automatically. It would monitor the status of all clients. Anyway, data is replicated so many times over so many computers that the statistical possibility of all the computers going down that have a specific piece of the puzzle is highly unlikely. Although, that possibility goes down as more clients volunteer.
----------------------------------------Why would they need ALL the data right away? We are always asking for updates on projects and we keep being told that they have so much data to go through that it will take time. Seems to me that they might not have to get to all the data at once. Either way, I didn't mean for this to become a debate. I was just suggesting an idea and we are discussing it. Seems feasible to me, but I understand the WCG people already have a lot on their plate. [Edit 1 times, last edit by Former Member at Jan 2, 2009 2:02:30 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Anyone frequenting the project statistics will have noticed that this project has been allowed allot more work units per day, 42 CPU years yesterday.
http://www.worldcommunitygrid.org/stat/viewProject.do?projectShortName=rice At fixed 10 hours they would be good to fill up the buffers more, so it would be advantages to do that and torque off the short HCC jobs to have less tasks in the pipeline come January 8. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yea, I'm running a temporary queue of 5.8 days worth of work to make sure I can make it over the hump without any idle time starting 1/8.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello Knightro2,
knreed posted that a 2 day buffer should be much more than needed. It is just that the possibility of immense unexpected server problems has to be gingerly disclosed to the WCG community at large. Since no practical size of buffer can compensate for the vast disasters that I can imagine (  I have lots of experience) I am keeping my buffer short.Lawrence |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yea, I'm running a temporary queue of 5.8 days worth of work to make sure I can make it over the hump without any idle time starting 1/8. Its great that you want to keep crunching through the downtime, really. But if everyone does that, we could introduce create an added headache. From the announcement... "we will start making changes soon to work unit deadlines so that we can reduce the amount of data to be transferred between hosting centers." I'm keeping all my settings the same, and if I don't work for a day or two, fine by me. They are expecting to lose a bit of processing over those few days. The point is the move will let them boost grid power in the long run. |
||
|
|
nasher
Veteran Cruncher USA Joined: Dec 2, 2005 Post Count: 1423 Status: Offline Project Badges: 
|
personaly i run a 0.5 days worth of work. the few days this project goes offline i will be crunchin other projects on the BOINC network... so i do not worry about the day or so downtime they will have (or even week or so)
----------------------------------------one problem by grabing LARGE amounts of work you also make a backlog of people waiting for there quorm of results to come back to get credit for the work they have done. the slowdown of the work units on rice dosnt seem too bad for me personaly though i get them fairly often right now   |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
So is RICE still slowed? Or did the researchers get the space they were needing?
|
||
|
|
|