| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Suggestions / Feedback
Thread: Grid computing for the masses - Bridge to other grids |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 5
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Since quite a time several different grids exists, each having significant computers count connected to them. It could be interesting idea to build kind of bridge to tap into resources offered by another grid (if the othr grid will allow it :-)
Seems like a new public domain grid is being built: http://machineslikeus.com/news/grid-computing-masses What is very interesting is the setup being very simple - so it might be interesting to start building this kind of bridge, that would consume tasks from World Community Grid and run them on the foreign grid. Another grid architecture could be XServe run by Apple, available for research and having a lot of potential... What do you think about it? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
And another powerfull GRID:
http://machineslikeus.com/news/grid-new-way-doing-science In Europe the data is carried over the GÉANT grid network but the organisation that makes this possible is managed by EGEE-III, the third phase of an EU-funded project to create an infrastructure supporting European researchers using grid computing resources. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
The BOINC client is that bridge. Attach to several projects, set the resource share and if any of these projects is out of commission, unreachable the client fetches or does work for other grids. In my case, work started coming in-trickle from Malaria.control.net and Docking@home, with a multi-week, very low resource share CPDN model.
----------------------------------------From client 6.10 you'll be able to set resource shares to zero for such backup projects. Only when there's really no work for WCG will the client then fetch work elsewhere and vice versa. Anyhow, when there's longer downs on e.g. SETI, we often see a little extra production on WCG. -^-//-^- PS: Think Almere grid (City in Holland) was set up with that thought in mind but never looked there. We've got current and future projects to last us till at least past the middle of the decade, taking growth into consideration, so no need for me to look hard at putting a larger share to other grid's disposal.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Feb 15, 2010 3:07:17 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
What you are saying is valid, but it applies to ONE computer only and it needs to have boinc configured and attached to WCG.
What I meant was fetching of massive work from World Community Grid and then distributing it on OTHER grid (not necessarily BOINC based) on all machines using the other grid (subscribed to the OTHER grid). This would be like a bridge between the grid servers, or a brigde appearing as one clinet on the SOURCE grid and then pushing tasks to the DESTINATION grid server of the other grid. In case the OTHER grid uses BOINC, it could be possible that the grids could migrate tasks between their projects, which is simple case. In case the other grid uses non-boinc environment (like the examples of grids I gave in this thread) then boinc-like execution environment must be simulated on the PCs running the tasks from the other grid. Now I understand that the technical challenge is big, but having big additional computational resources is tempting, is not it? Hope I was clearer this time ... |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
No, it applies to all computers since resource shares are controlled from the web profiles (the 4 venues) and once it is applied, it propagates automatically to the clients, all clients on an account/venue.
----------------------------------------I did not mention GridRepublic or BAM, but's also kind of a multi-project coordinator... you can mass attach/detach using their web tools and change the different grid resource shares, WCG still working as the last to become fully integrated (which requires at least BAM too to make changes). Resources needed to do that: Fork over lots of dosh. None BOINC clients support or emulation, testing, debugging... think that's several bridges too far. I do understand what you're saying, but it adds massive complexity in my opinion.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
|