| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Discovering Dengue Drugs - Together
Thread: Server-side co-location.. a defense against future project upsets due to extreme weather / disasters? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 5
|
|
| Author |
|
|
AppleTV
Cruncher Joined: Dec 7, 2008 Post Count: 45 Status: Offline |
On a side note, is there a possibility that the project could be co-located to protect against likely future weather mishaps such as what caused this down time? If I believe that climate change is a real problem, then I also have to accept that we need to be more prepared for more of this nonsense in greater frequency and scales of damage in the future. It makes me deeply worried to observe that there seems to be no great effort underway to co-locate projects such as this as a defense against things like natural disasters which should not be allowed to terminate or suspend important research...
I am sitting here looking at the map and thinking to myself, the same thing that took out this project could happen just by a small tsunami and wipe out all the research going on in the SF Bay Area (Silicon Valley) just as easily. As a society, I think it's time to accept these dangers as actualities and start enacting plans to make critical data and important scientific research less location-dependent and more decentralized. Is there a way to do this within grid computing? Granted the client side is as decentralized and protected as could ever be. Could there also be a way to decentralize to some degree the server side so that the loss of one location does not become the loss of the entire project? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
That word doesn't mean what you think it does. Colocation is where multiple servers (usually from different companies) are located in a single data center.
WCG run their servers in a secure IBM data center. Individual projects have their own IT solutions. Obviously, you are talking about decentralisation. I have to say, the additional costs aren't justified. There is little enough money for research as it is. Finally: the problems experienced by DDDT are only partly technical. I imagine that all the data is safely backed up, so even in the event of a cataclysmic disaster, it could all be recovered. Relatively, though, the human resources are more valuable. |
||
|
|
AppleTV
Cruncher Joined: Dec 7, 2008 Post Count: 45 Status: Offline |
That word doesn't mean what you think it does. Colocation is where multiple servers (usually from different companies) are located in a single data center. WCG run their servers in a secure IBM data center. Individual projects have their own IT solutions. Obviously, you are talking about decentralisation. I have to say, the additional costs aren't justified. There is little enough money for research as it is. Finally: the problems experienced by DDDT are only partly technical. I imagine that all the data is safely backed up, so even in the event of a cataclysmic disaster, it could all be recovered. Relatively, though, the human resources are more valuable. Help me understand then how a hurricane in Texas had any impact on this project if IBM's data center was the issue? I thought IBM's facilities were concentrated in the Northeast somewhere. You are right. Colocation wasn't the right word, but by colocation, one can also infer moving or backing up data from risky locations to more secure locations, as companies do with colocation facilities. Companies like Google keep copies of millions of TeraBytes of people's private data online just for their "free" gmail service. One would think that eventually they would step up and create free remote backup mirroring solutions for scientific applications as well. Even if it's not a breadwinner, it's certainly good PR.. Although economics should not be a motive for doing things like supporting scientific research even some of the research benefits them if in the indirect. I imagine a very successful CEP would help support Google powering their servers more efficiently from the sun instead of tapping the power grid or overpriced contemporary silicon panels. And then there are executives with sons, daughters, sisters, brothers who may have diseases targeted by these projects. Sometimes economics is not the best motive for scientific research. It's just one motive. IBM doesn't necessarily make money on WCG does it? Surely other corporate sponsors can be leveraged to help provide infrastructure restoration or duplicity of resources so that geographic-specific events can't grind research to a halt again like this? That's what I'm trying to get at here.. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
IBM spend money on World Community Grid. It's entirely not for profit.
What I was trying to say is that it wasn't particularly due to the hardware that DDDT had problems. Most of the downtime has been due to the researchers and students trying to put their lives back together, find accommodation, stuff like that. The data was safe. The researchers put together their own IT infrastructure. They know what they need. Often they are going to need direct access to the data for post-processing. IBM can (and do) help with grants. Obviously, it would be great if Google or other companies helped, too. |
||
|
|
AppleTV
Cruncher Joined: Dec 7, 2008 Post Count: 45 Status: Offline |
IBM spend money on World Community Grid. It's entirely not for profit. What I was trying to say is that it wasn't particularly due to the hardware that DDDT had problems. Most of the downtime has been due to the researchers and students trying to put their lives back together, find accommodation, stuff like that. The data was safe. Ah, OK, now I see. I thought it was a technical issue due to all the talk of flooded facilities. |
||
|
|
|