Universities, government labs, and sometimes IT vendors donate their excess supercomputing capacity through grants to academics to help advance various sciences. Now Google is letting boffins loose on its systems. In a blog post, Alfred Spector, vice president of research and special initiatives at the Chocolate Factory, said …
Fair play to them
That's a nice gesture
No doubt there will be a lot of cynicism in here any minute now, but this isn't something that they 'have' to do, so props!
Oh go on then
Not really meant as cynicism but they could have thrown their lot in with the World Community Grid who do this sort of thing already. They would still have got all the publicity although it would have meant working with IBM so maybe that put them off.
World Community Grid
I was thinking that they could just write a client for the World Community Grid to run on their system as well, but I wonder if the Academic Grant is being used as a shake down run before they charge businesses for the same service and as such the donation of cycles isn't actually the aim of the project.
So who owns the research?
"Those who win the grants will do their work from Google offices and sign an employee agreement with Google for the term of the simulation"
This suggests again that Google is looking to amass information and rights to research. I wonder what the long term view of this research and those who depend on it will be?.. Are they donating or are they in fact looking to "Slurp" the research under an employee agreement then sell the rights to it later...
Yes Cynical ... but it is Google... they dont do anything for FREE.. there has to be an angle either visible now or further down the line.
Cynical it may be but how many people will clamour to ge tthe computer hours and sign a contract with the devil for the devil to turn round and say.."You were an employee... all the work you did is ours.."
Let's read that a little closer...
"And the researcher will have to be at a Google facility AND sign an employee agreement"
If anyone bothers to read employment agreements:
Anything you discover, create, etc. while under a standard employee agreement is the property of the employer, who can (and usually does) grant essentially nothing (<<1%) to the employee.
By working in and on Goolge equipment and facilities Google further creates a condition of ownership over any intellectual property created while using the Google facilities.
It's going to take some really sharp attorneys to thresh out all of the hidden consequences in this deal.
Note - Shame on the Reg for not catching this opportunity for Google to accidentally claim ownership over any significant new discoveries.
What really offends me is that Google does not even have the decency to pay travel, lodging and a stipend to the researches whose work it is setting up to be hijacked. On the other hand by failing to "consummate" the employee status it has left a tiny legal door open for the researchers try and break the employee agreement.
This whole "offer" looks like a legal nightmare brewing with possible tax consequences since providing research to "employees" is not the same as donating grants to non-profit universities.
The "grant" arrangement just does not look as smart as what I expect from Google (except for the opportunity to appropriate others research). Maybe their recent leadership change is grounds to ditch their stock.
oh no, they have to pay to fulfill a lifelong dream
Most likely they will receive funding from whichever institution they work with. Most of the people likely to be going for this kind of deal will be academics and their institutions will most likely be very willing to pay for the possibility of a great discovery being associated with them.
In terms of what google will own of the research that is yet to be known. These are smart people who will probably read the contract quite a few times before they make a decision. If they feel google takes too much then they wont do it. On the other hand they may feel its worth the cost to get their work realised.
Not sure on the details
I'm sure I read somewhere that if you create some patentable invention/creation (genuine are few and far between these days) then, no matter what your employment agreement states, you are entitled to a reasonable return on the licensing. The employer can state that they own the copyright but you are entitled to a cut of patent licensing. I think this may have related to the UK rather than the US but I'd be interested if someone knows more details.
Depends on what jurisdiction you're in. Different countries have different laws. If you're in the US, then different states probably have different laws. Consult a local lawyer.
When I've invented things, my employer has paid me a bonus of around £1000 for each patent application. This is a one-off, up-front payement; I've signed away my rights to any ongoing royalty. (And I get paid even if the patent offices reject the patent).
a unique volunteer computing grid ?
Isn't that what Boinc is ? With more than 300,000 volunteers ?
Are you suggesting that Google "stole" your idea ? Because if so, it seems to me that you're late to the party.
Google, install BOINC on your Servers!!!
UTTER AND MAJOR FAIL.
If they want to donate their cycles to science, (like I do) they should access the aforementioned site (and I repeat it here: http://boinc.berkeley.edu/ ) and start downloading already registered science projects (work units). They just have to be sure they can cope with 100% CPU usage before they get it properly configured, should they choose to do so.
No dodgy copyright issues. No draconian EULAS. Choose a project and run it. Given the might of the Google servers, I would promptly suggest climateprediction.net since their Work Units take 700 hours on a Intel Core i7 home PC running at full tilt.
Just run its projects at 1 work unit per core and watch all the projects ticking "completed" before you can say "Google is not evil" 1000 times.
Anonymous Coward, no doubt Google is grateful for your suggestion
that the firm access the BOINC site ; Google, however, does seem to have been aware of the latter even previously - note the following passage under the subheading Technical Specifications and Requirements on the web page which explains the Google Exacycle for Visiting Faculty Grant Program (http://research.google.com/university/exacycle_program.html) :
«Proposals that are ideal for Google Exacycle include, but are not limited to, research projects like Folding@Home, Rosetta@Home, various BOINC projects, and grid parameter sweeps. Other examples include large-scale genomic search and alignment, protein family modeling and sky survey image analysis.»
Always wise to do one's research before firing off a comment....
The Joke Icon is there for a reason. No misplacement of FAIL icon.
Still, the fact that you can borrow Google's cycles is very nice. Of course they would have read BOINC projects before opening this kind of venture to the public, otherwise they've kept their mouths (mice?) shut and start running these already BOINC'fied projects.
My bad, Anonymous Coward -
I had thought that jokes were supposed to be funny....
Grousing about Google
Only came in here to find out how somebody could find a way to complain about Google even in this. Mission accomplished.
Not So Useful
Moving from the legal to the technical aspects of this offer... the lack of decent interconnect will make it useful for only a small fraction of supercomputing tasks. Anyone wishing to do complex physics simulations will be out of luck.
@John Greenshields, @oliver 8, @AC 16:07 GMT et al
Ok, this time without the abuse which got my last post canned and perhaps rightly so.
If you bothered following the link in the article you'd find this
"At completion of the project, we recommend, but do not require, that all the researcher's data be made freely available to the academic community. "
Ok, took under 30 seconds to find that. So next time try either checking the facts or not assuming the worst of everybody. Idle cynicism unbecomes you.
Don't Google know about this great wired world ?
"You have to pay your own travel, lodging and living expenses while the simulations run,"
Why does the user have to be there, wherever there is ?
Maybe because they don't want to open up their network for remoting from someone whose machine they do not control?
This pleases me
It's nice to know this sort of computing power is up for grabs if something really important requires it.
What would be utterly magnificent is if we could get an annual 'No Google Day', in which ALL their resources were pointed at something like World Community Grid for twenty-four hours.
I guess the businesses that rely on Google Docs and suchlike wouldn't be too pleased but... um, I dunno, make it on Christmas or something.
...just spread the resources that that would have freed up across the other 364 days of the year, supplying the same benefit (good) without inconveniencing your paying customers (evil)
It pleases me too, but it's sad to see that it's possible to be called wicked-bad when doing good. I can imagine some people here being given a free lolipop and throwing it back in the kind donor's face yelling "This is blackcurrant flavour! I wanted strawberry! You're EVIL, free lolipop person! Nyaah!"
Google has millions of servers
Does Google really have millions of servers? What evidence is there to support this? I know they have a fair few but realistically even 1000000 servers is a lot.
You also quote 36 data centers, this would average 27777 servers per DC, requiring 661 42u racks per DC ignoring the fact that google are almost certainly not using 1u servers and that they would not put 42 servers in a rack. More likely they will use 4u servers since they are widely known to use off the shelf components. Add in some (LOTS of) network equipment and replicated SAN etc and this would be a very large datacenter indeed given the dimensions of a standard rack.
I'm not saying you're wrong but I don't believe that Google do have millions of servers and I don't think they need that much processing power either.
Somewhere to compile my new Pong game