Thread 'NSF funds new model for BOINC'

Message boards : News : NSF funds new model for BOINC
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · 4 · Next

AuthorMessage
ProfileDavid Anderson
Volunteer moderator
Project administrator
Project developer
Avatar

Send message
Joined: 10 Sep 05
Posts: 726
Message 78714 - Posted: 6 Jun 2017, 21:15:59 UTC
Last modified: 6 Jun 2017, 21:16:30 UTC

The National Science Foundation has funded a 3-year project by UC Berkeley, Purdue University, and Texas Advanced Computing Center to develop a new framework for BOINC-based volunteer computing, in which volunteers sign up for science goals rather than projects. Details are here.
ID: 78714 · Report as offensive
Egilman
Avatar

Send message
Joined: 27 Jan 16
Posts: 16
Message 78722 - Posted: 7 Jun 2017, 2:28:12 UTC - in response to Message 78714.  
Last modified: 7 Jun 2017, 3:17:53 UTC

Dr Anderson,

Personally, I think your developing your way out of volunteer computing.

Everyone I've had a chance to talk to about your TBD posting "http://boinc.berkeley.edu/tbd.php" of the 1st states unequivocally that in no way would they remain involved if you took their project/work choice out of the equation.

Neither would I.

Just my opinion but I am the #1 non-asic boinc producer in the world. We exchanged a few emails back in the old Seti I days concerning the last cheating scandal and the future. You iterated for me your vision of what Boinc was to be, and it was good. Turned me into a believer.

Your "TBD" proposal just tells me that DC has run it's course on a volunteer basis.

If it does come to fruition, I will not be along for the ride.....

Thank you for the 15 years of fun....

EG, formerly of Ars Technica, currently of Overclock.net

I sincerely wish you good luck with your proposal... (I just don't see how it is going to work for us individuals out here)
ID: 78722 · Report as offensive
No.15

Send message
Joined: 7 Jun 17
Posts: 1
United States
Message 78724 - Posted: 7 Jun 2017, 4:05:32 UTC - in response to Message 78722.  

I agree with the Egilman. TBD as a project is not something I would participate in as you have explained it. To me it looks as a way to monetize the projects with the end user never really knowing what they are crunching and who they are crunching it for. My prediction would be that user numbers would drop because of this.

In the forums I participate in, there has not been 1 post thinking this is a good decision.

Perhaps you could explain to me, an average cruncher, why I would want to participate in TBD instead of Boinc?
ID: 78724 · Report as offensive
Qax

Send message
Joined: 7 Jun 17
Posts: 1
United States
Message 78725 - Posted: 7 Jun 2017, 4:33:04 UTC - in response to Message 78714.  

I agree with these guys. People want to choose where their resources are going.....people who boinc still. We like the current system.

If there was a group "SPACE" - I mean....there are so many different projects and goals that that could be. Already within programs some already split resources (cern has multiple projects, some projects have beta testing, SETI has had astropulse and other, and the World Community Grid has many different goals).

People donate money to charities specifically -- they don't just sent their money away to a consortium of all of the characterizes dealing with cancer or hunger, or human rights and let them split up our resources. This plan is completely deaf to what makes BOINCing fun for us.

There do need to be some changes made -- we do need more users and to stop losing the ones we do have from attrition. This is not the solution though to that problem. I have some ideas.....but they are very different than this. I did see a problem and bought the domain boinc.org, hoping to eventually make a non-profit to try to keep people around longer.

Whatever is being created, I am sure I will be out at that point.
ID: 78725 · Report as offensive
Bernie Vine
Volunteer moderator
Avatar

Send message
Joined: 10 Dec 12
Posts: 323
Message 78731 - Posted: 7 Jun 2017, 7:15:50 UTC

The way I read it is this will work alongside BOINC

I suspect that relatively few current volunteers will use TBD; it's more for new users with wider demographics. So current projects won't lose computing power,


and

To implement TBD, I'll need to add some features to BOINC, e.g.:

The client will pass credit estimate information to account managers.
Account managers can send clients opaque data to be passed in scheduler requests (preference keywords in this case).
The scheduler will have a "keyword matching" option that takes user and job keywords into account. E.g. it will preferentially send biomed jobs to volunteers who want to support biomed.
These features will have no impact on existing projects.


(emphasis is mine)

I suppose I could be wrong.
ID: 78731 · Report as offensive
Sirius B
Avatar

Send message
Joined: 12 Jun 09
Posts: 2102
Ireland
Message 78735 - Posted: 7 Jun 2017, 7:44:13 UTC

I could be wrong also, but after reading about TBD, have some concerns: - 10's & 10's of millions & corporate partners? I see it as a corporate gold rush. What a data mining freebie.
ID: 78735 · Report as offensive
boboviz
Help desk expert

Send message
Joined: 12 Feb 11
Posts: 418
Italy
Message 78746 - Posted: 7 Jun 2017, 9:14:12 UTC

I continue to remain with OLD boinc, but some points
Creating and operating a VC project is harder than we realized: it requires a combination of resource and skills (Win/Mac programming, sysadmin, DB admin, web design, PR/outreach) that few academic research groups have.

It's your fault. With good and continously updated documentation, support on forum/mailing list/etc, the administration of a VC project for a research group is NOT so hard.
If academic group expects everything on a plate, maybe they don't want to do research seriously.

For a research group, trying to use VC is a risk. There's a substantial investment, with no guarantee of any return, since no one may volunteer.
Adding a VC component to a grant proposal adds uncertainty and weakens the proposal.

Investment? You can run your preconfigured boinc server virtual machine on a simple workstation to start a new project. Is this an investment?
For volunteers, it's not the first project that close because they have MORE computational power than they expected!!

The computing needs of many research groups are sporadic - e.g. they need a big chunk of throughput every now and then. For such groups, buying computing time on a commercial cloud may be cheaper than using VC.

I cannot understand. They have money to buy (Amazon, Azure, etc) cloud but they have no money to invest (see above) in boinc server?
And i don't think that your TBD will give the "continuity of service" of an HPC infrastructure. And this is NOT THE INTENTION of VC.

Attracting volunteers is a marketing exercise. It's difficult to do effective marketing when there are dozens of competing brands (i.e. projects names).

No, it's simple. If you, for example, have "respect" for volunteers (participation on forum, publication of news, etc.) you will have a lot of volunteers and computational power.

Since then, nothing: no conferences on distributed computing list VC as a topic of interest.

During my research for boinc publications, i found a lot of papers, posters, phd thesis about boinc. And also conferences where boinc is discussed.
ID: 78746 · Report as offensive
mmonnin

Send message
Joined: 1 Jul 16
Posts: 146
United States
Message 78751 - Posted: 7 Jun 2017, 10:24:29 UTC

David, After replying to you that I couldn't find your link anywhere I posted on my home forum and it started a lively conversation. Here are the comments I posted there:

I don't think there will be extra millions of people that will run this model. DC/VC/FAH/BOINC costs electricity to run and at 100% processor utilization many Dell/HPs/etc fans start to get loud. With those two things you're just be left with the dedicated who are already running BOINC as is. Probably some more but I'm not sure millions more.

'For the Science' is nice and all but let's be honest. It's a competition. There are points involved. Some homogeneous blend of points across the group of projects won't be fun. And it won't be easy to get work from one project to give out the same points as another. Some projects get more or less users due to their high or low PPD output as is. That's a clear indication of the value of points and competition.

Control. I'm guessing we'd lose control over exactly what we run and are left to groups like math, astronomy, medical. Some projects stress hardware much more than others. We'd have to tune for the hardest and lose efficiency on another. Then there's the memory/disk requirements as well. How does one configure it like we can BOINC now.

I do see the benefit of a generic computing grid where smaller projects with intermittent work can get the computing cycles they need.


I've seen it over and over again through the years. People get their 1st electricity bill and they are done with any type of DC (VC). Money talks. The millions won't come flocking because its easier.

The benefit here might be for the small groups that need some instant compute power. But if projects start going to TBD instead of BOINC cause its easier to setup and don't have to deal with the people supplying their power, aka everyone on this forum, then its the death of BOINC's variety. The widespread choices of what to crunch and when I want to crunch it will be more limited. Thats the great think about BOINC vs FAH. Folding is the same thing all day, every day and thats what TBD will be. Boring.

Since it already seems like a sure thing since it has funding, if you want people to donate hardware and their power bill, you'll have to make it interesting. Points, badges, transparency.
ID: 78751 · Report as offensive
Bryan

Send message
Joined: 7 Jun 17
Posts: 1
United States
Message 78766 - Posted: 7 Jun 2017, 16:08:09 UTC

Gentlemen, you are all missing the point of this!

A while ago NSF dropped their grant (money) to fund BOINC. DA in effect lost his job. He immediately left his baby, BOINC, and put future development/support into the private sector. He washed his hands because he was no longer being paid.

So what do you do if you want to again feed from the NSF money cow? Of course you come up with another idea and sell it to them! You sell them on the idea that millions and millions of people will connect their phones, tablets, and computers to help major data centers crunch for science. Viola, you get a job and get paid for another 3 years ... now that is success! Of course it is a shame that the other 2 collaborators only get 50% of a salary, but thank goodness DA is taken care of.

And now that he is getting paid again, he will come help us with BOINC. I feel so blessed. Wait a minute, didn't he say he was coming to help BOINC because he needed something from it to support his new TBD project. Nah, that can't be it. He's coming to take care of his baby again because he cares!

Sorry for taking to the soapbox. I'm old and cynical :)
ID: 78766 · Report as offensive
ProfileGary Charpentier
Avatar

Send message
Joined: 23 Feb 08
Posts: 2486
United States
Message 78768 - Posted: 7 Jun 2017, 16:55:48 UTC - in response to Message 78714.  

Unless you have done some focus groups on BOINC users you may be reading the reasons far wrong as to why it didn't take off. You may have the part on why few projects are offered somewhat better.

I suspect the reason is more that the projects offered are not sexy and none have had a spectacular success with the donated computing. As you say you aren't a marketing type and I'm sure other scientists and researchers aren't either. Also there is that research for profit angle. Why donate $$, which you can't even get a tax deduction for, to some mega-corp to make a pile of cash?

However as desktops are gone and phones will have no spare cycles to donate, twitter, facebook, youtube and their ilk will see to that and eat all bandwidth for file transfers, the time of donated computing is over.

The only reason a couple of the projects have any computation at all is because they pay credits much larger than credit screw does, and yes there are people vain enough to pay to have a big number next to their name on a website.
ID: 78768 · Report as offensive
ProfileDavid Anderson
Volunteer moderator
Project administrator
Project developer
Avatar

Send message
Joined: 10 Sep 05
Posts: 726
Message 78776 - Posted: 7 Jun 2017, 20:56:28 UTC

Interesting reactions, and some misconceptions. In no particular order:

- There are and will continue to be plenty of powerful desktop/laptop computers.
- Smart phones have plenty of free cycles and free WiFi network capacity.
- BOINC projects have had, and continue to have, research results published in top journals. Spectacular? Matter of opinion.
- Few if any BOINC projects are "research for profit".
- TBD is an account manager, not a project. If we succeed there will be more projects and more variety.
- "Corporate partnerships" are things like Intel Progress Thru Processors, HTC Power to Give, etc. The corporations promote VC; they don't make any money from it.
- To repeat: TBD doesn't preclude anything. TBD lets people contribute to science areas rather than specific projects. People who want to participate in specific projects can continue to do so.
- We've studied volunteer motivations extensively; see the various papers by Oded Nov. More people are motivated by science goals than by credit.

-- David
ID: 78776 · Report as offensive
Egilman
Avatar

Send message
Joined: 27 Jan 16
Posts: 16
Message 78779 - Posted: 7 Jun 2017, 21:20:32 UTC - in response to Message 78776.  
Last modified: 7 Jun 2017, 21:28:18 UTC

Good points David.

But let me point out something.

There are now over 2 billion android devices in the wild, in over 24k flavors. ALL of them connected 24/7, ALL of them doing background tasks even when turned off.

Android devices are very capable of doing distributed work, Boinc has proven that beyond any doubt. Both standard processing and parallel processing.

Sounds like your stating that the failure of Boinc is the user having to choose what work gets done. which leave some projects with an abundance of processing power and others starving for power.

TBD as described seems to be aimed at those social network devices that operate in the background, connected and communicating to the net, without the user knowing that they are still operating. That is the way they are designed.

So essentially your turning the social network system into it's own supercomputer to be used for whatever the desires of those running the net want it to do....

And the owner of the device has no choice..... to not participate he has to disable his device in a way that prevents it from communicating to the network. Which they are designed to NOT do.

You seem to be creating a system that is capable of eliminating a users choice altogether. Which is what you seem to view as the failure of Boinc.

I don't own an android device, cause a blind man could see this coming.

I will not allow my devices to be used for something I do not know what....

Just my feelings on the matter.

Unless I'm completely wrong....
ID: 78779 · Report as offensive
mmonnin

Send message
Joined: 1 Jul 16
Posts: 146
United States
Message 78780 - Posted: 7 Jun 2017, 21:48:10 UTC
Last modified: 7 Jun 2017, 21:48:48 UTC

I'll have to read through some of these links if the documentation is available. Do any of them say that people are more motivated by goals vs credit if and ONLY if they have to pay for it? IE with hardware and electricity? Do these motivated people know their mobile device will be tethered to the wall if they want to contribute? The Sony FAH app only counts by hours and it doesn't take long to get into the top several % so I don't think the millions of people are there.

http://research.poly.edu/~onov/
ID: 78780 · Report as offensive
Jim1348

Send message
Joined: 8 Nov 10
Posts: 310
United States
Message 78781 - Posted: 7 Jun 2017, 21:55:31 UTC
Last modified: 7 Jun 2017, 22:17:36 UTC

I think it is an outstanding idea. If the present BOINC population isn't interested, that is largely irrelevant. They are not the target audience. Most people you would want to attract (assuming they are there at all) will find "BioMed" or "Astronomy" or whatever to be sufficiently descriptive. They are not the people who hover over their machines most of the day anyway.

I have been concerned for some time that as computing power increases, more projects will go in-house. Each time I hear a whine session that some BOINC project is not supporting an old video card, CPU or whatever, I think that the trouble managing BOINC may be getting greater than the benefits. That is partly the fault of the administrators in trying to cater to everyone's whims, but it seems to be solved definitively in the new model insofar as I can tell. The devices either work or they don't, and the target audience will not be spending much of their (or the administrator's) time trying to hassle it out, however amusing that might be.

So I don't see that anything is lost, and maybe a lot to be gained. It is worth a try.

By the way, I am all in favor of supporting quasi-commercial research, if that happens to be included. That is much closer to real-world applications. This notion that if no one can profit from it then it must be good research comes from people who are naive in the field. There may need to be some required licensing provisions or whatever so that companies are playing on a level field, but you need to make any research profitable at some point for it to be used.
ID: 78781 · Report as offensive
Egilman
Avatar

Send message
Joined: 27 Jan 16
Posts: 16
Message 78782 - Posted: 7 Jun 2017, 22:30:07 UTC - in response to Message 78776.  
Last modified: 7 Jun 2017, 22:31:28 UTC


.......

- We've studied volunteer motivations extensively; see the various papers by Oded Nov. More people are motivated by science goals than by credit.

-- David


Yes you've written two papers in collaboration with him and others, or maybe it was just a contribution to a paper...

Volunteer computing: a model of the factors determining contribution to community-based scientific research http://dl.acm.org/citation.cfm?doid=1772690.1772766 Published from WWW '10 Proceedings of the 19th international conference on World wide web conference, April 26th 2010

And

Dusting for science: motivation and participation of digital citizen science volunteers http://dl.acm.org/citation.cfm?doid=1940761.1940771 Published from iConference '11 Proceedings of the 2011 iConference, February 11th 2011 (I attended this conference)

He's wrote a few other papers also on IT system design and human behaviors.

I take it you believe as he does.....

Just so you know I've been following your publications since your first in 2001 with Dan, Jeff, Matt & Eric....

So your turn to this type of model is kinda a surprise for me....

Just kinda trying to understand why your turning away from the open volunteer model?
ID: 78782 · Report as offensive
zioriga

Send message
Joined: 27 Nov 14
Posts: 12
Italy
Message 78784 - Posted: 8 Jun 2017, 4:47:05 UTC

just some questions:

TDB and standard BOINC are mutually exclusive ?? or they can run simultaneously on a single computer ??

If I'll participate, for example, in the TDB Cosmology field (or in any other field) there will be a way to know which computing goal i'm working for ?? Or will be possible to know something about the science behind every working unit (or most working units) ??
ID: 78784 · Report as offensive
noderaser
Avatar

Send message
Joined: 2 Jan 14
Posts: 276
United States
Message 78785 - Posted: 8 Jun 2017, 5:25:39 UTC

Sounds like an interesting idea, I think perhaps that simplifying the way users interact with their "VC" application would make it easier for new people to get involved. Even as a long-time contributor, I find the aspect of managing the projects and preferences can be somewhat tedious and time-consuming--and know that it could be a turn-off to new users who are less technically inclined or invested in their contributions. Obviously, the regular BOINC system should remain in place, but a new account manager where you pick general areas of research to contribute to really resonates with me, you don't have to worry about the minutiae of going through the project list, creating accounts, etc. If it could be paired with a very simple app, there's a chance we could get more people involved. Just install the software/app, pick your areas of focus, and be off.

I would prefer to see that power used to first support the existing projects, then expand to other endeavors. I know there are some users who won't touch anything remotely related to for-profit work, so that should be an important opt-in function.

But, the whole idea of VC is pretty techy-nerdy, I don't foresee it getting much farther than BOINC unless there are some marketing geniuses involved.
My Detailed BOINC Stats
ID: 78785 · Report as offensive
boboviz
Help desk expert

Send message
Joined: 12 Feb 11
Posts: 418
Italy
Message 78791 - Posted: 8 Jun 2017, 8:14:05 UTC - in response to Message 78766.  

A while ago NSF dropped their grant (money) to fund BOINC. DA in effect lost his job. He immediately left his baby, BOINC, and put future development/support into the private sector. He washed his hands because he was no longer being paid.


You are wrong. For DA is not full time work, but is continuing to work on boinc platform (https://github.com/BOINC/boinc/commits/master)
ID: 78791 · Report as offensive
boboviz
Help desk expert

Send message
Joined: 12 Feb 11
Posts: 418
Italy
Message 78792 - Posted: 8 Jun 2017, 8:15:13 UTC - in response to Message 78781.  

By the way, I am all in favor of supporting quasi-commercial research, if that happens to be included. That is much closer to real-world applications. This notion that if no one can profit from it then it must be good research comes from people who are naive in the field.


Why? Why is much "closer to real-world"??
ID: 78792 · Report as offensive
robsmith
Volunteer tester
Help desk expert

Send message
Joined: 25 May 09
Posts: 1295
United Kingdom
Message 78794 - Posted: 8 Jun 2017, 8:41:53 UTC

Regardless of what platform (BOINC/TBA/Whatever) volunteer participation will depend on the job in hand being "attractive" to the potential volunteer. Over the years some projects have been "attractive" and have thus attracted a large user base, while other projects have not had that element of attraction and have struggled to gain the required volunteer base.
I don't think anything can be done by the platform to solve this issue, it is very much down the project being "attractive", marketed properly and managed properly.

As to the matter of the back-office required by the platform on the project - David has identified that as being an issue for many smaller projects in his paper, and unless there is a substantial change in the underlying philosophy I can't see that changing significantly. There are certain functions that don't change, e.g. user management, application development, data management.
ID: 78794 · Report as offensive
1 · 2 · 3 · 4 · Next

Message boards : News : NSF funds new model for BOINC

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.