[kepler-dev] Question regarding distributed kepler

Jianwu Wang jianwu at sdsc.edu
Mon Sep 8 11:08:21 PDT 2008


Hi Yaser,

    Thanks for your interesting on distributed kepler. I'd love to join 
the conference call. 

    As Derik mentioned, there is a interest group on distributed 
execution 
(https://dev.kepler-project.org/developers/interest-groups/distributed).  
Derik, Chad and I are main members. It would be great if we can 
collaborate on distributed  Kepler. 

    If you have concrete requirement on it, would you please tell us 
about it?

Best wishes

Sincerely yours

Jianwu Wang
jianwu at sdsc.edu

Post-Doctor
Scientific Workflow Automation Technologies (SWAT) Laboratory
San Diego Supercomputer Center 
University of California, San Diego
San Diego, U.S.A. 



Derik Barseghian wrote:
> Hi Yaser,
>
> On Sep 6, 2008, at 1:23 PM, Yaser Jararweh wrote:
>
>> Dear Derik,
>>
>> I hope my email find you doing vrey well,
>>
>> FYI, we are now using the SVN version of kepler and we are now able 
>> to write our own Actors from scratch, also we connect Kepler with 
>> DataTurbin by building an interface between both of them to read and 
>> write the data streaming.
>>
>
> Interesting. I am curious to know the details of what you have 
> designed and what datastreams you are dealing with.
> And were you able to make use of the DataTurbine actor?
>
>> We are now working to distribute our models over a cluster of 
>> computing nodes, and we have some question regarding this issue.
>>
>> 1- Can we write and compile our own director (customise director).
>
> Yes. Please see this thread for details, eventually following the link 
> to Edward's presentations about directors:
> http://mercury.nceas.ucsb.edu/ecoinformatics/pipermail/kepler-dev/2008-July/006477.html 
>
>
>> 2- Can we have a a director that control other directors from 
>> different types.
>
> Yes. Please see this paper:
> http://www.eecs.berkeley.edu/Pubs/TechRpts/2007/EECS-2007-139.pdf
>
> Also, this thread discusses contained directors a bit:
> http://tech.groups.yahoo.com/group/ptolemy-hackers/message/1741?threaded=1 
>
>
>
>> 3- we noted that u have something regard the distributed Kepler using 
>> (RMI) but there is no real implementation for it, can you explain for 
>> us what are you planning to do regarding that issues.
>>
>
> Did you see the webpage about this?
> http://www.kepler-project.org/Wiki.jsp?page=InstallingDistributedWorkflows 
>
>
> Chad Berkley has been working on this. He may be able to give more 
> details.
>
>> Also can we have a conference call with you to talk regard this issues.
>>
>>
>
> Yes, maybe near the end of this week or next week (I'm leaving for the 
> EIM conference tomorrow). I haven't worked on distributed computing in 
> Kepler, so I will see if others who have could join a call. I will be 
> working on this in the future however. We are redesigning the Kepler 
> website, and it has a section for distributed computing that we hope 
> will be useful to the community for communication and collaboration 
> around this topic:
> https://dev.kepler-project.org/developers/interest-groups/distributed
>
>>
>> -- 
>> Thank You,
>>
>> Yaser
>
>
>
> Derik
> _______________________________________________
> Kepler-dev mailing list
> Kepler-dev at ecoinformatics.org
> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-dev


More information about the Kepler-dev mailing list