[kepler-users] WSDL interface

[Stuart Chalk]

Matt, Paul

I guess it would help if I explained what I wanted to do and why  
kepler with a wsdl interface would be good.

I am trying to write some software to collect, organize, analyze, and  
disseminate science data (I am an analytical chemist - CSI lab guy;)).
This project is up at http://eureka.sourceforge.net although the  
material I have there suggests the software be done in cakephp  
exclusively.

The setup I want to use is cakephp as the "engine" for the software,  
data would be stored in Fedora (http://www.fedora-commons.org) as XML  
files,
and workflows would be made in kepler and stored in Fedora for future  
use.  I need a mechanism for the cakephp engine to get the data to  
process
(from fedora), grab the correct workflow (from fedora), and send the  
data for processing.

Getting the processing to happen does not have to be done using wsdl  
interface, I just thought it made sense to do it that way.
If I can use cakephp to call kepler from the command line (i.e.  
running locally on a mac) then that is fine.
To me I would like to keep as much of the software platform neutral as  
possible and thus wsdl looked better than command line.

If a wsdl interface is a way off, or impractical then I'll use the  
command line approach for now.

Stuart Chalk

On Feb 18, 2008, at 1:15 PM, Matthew Jones wrote:

> Hi Paul and Stuart,
>
> I'm not quite sure what you are trying to accomplish with a web  
> services interface, but I think it is probably clode to what we were  
> aiming for with our distributed execution feature.
>
> The distributed execution features you describe are working in the  
> Kepler head, but only the basic framework is in place.  You can  
> define a workflow, and it can be run on a set of Kepler slave nodes  
> that the user can configure.  There is still a lot of testing  
> needed, and a number of additional user-interface features are  
> needed to make it easy to deploy and use.  Lucas could probably give  
> you an update, but I think the second web page that Paul cited is  
> current (Chad updated it recently).
>
> The distributed execution feature allows one to run sub-workflows  
> (defined using the Discributed COmposite Actor) on remote nodes, and  
> the Kepler system takes care of moving the required workflow  
> specification and data to the remote node, executing the code, and  
> returning the results. This is somewhat different than exposing a  
> workflow as a web service that can be called from other clients.
>
> The current distributed execution implementation takes care of data  
> staging, so unlike in Paul's case, the data is returned to the  
> caller as specified in the workflow specification.  I suspect a web  
> services layer could be placed in front of the distributed nodes,  
> but maybe this isn't needed if the distributed execution feature is  
> really what is needed. Maybe Chad or Lucas would have more  
> information on whether this would be possible or desirable.
>
> Regards,
> Matt
>
> Paul Allen wrote:
>> There is some work and design that has been done for executing  
>> workflows in a distributed fashion, but I'm not sure of their status.
>> See http://www.kepler-project.org/Wiki.jsp?page=DistributedKepler  
>> and http://www.kepler-project.org/Wiki.jsp?page=WorkingDistributedFeatures
>> I am building a system to have a stable of (remote) Kepler engines  
>> watching a job queue and then picking jobs to run. It wouldn't be  
>> hard for me to add a web service interface to the job queue  
>> section. However, you'll find that one of the challenging issues  
>> about remote execution, is that the workflow has to be smart about  
>> where to put its results.
>> -Paul
>> Stuart Chalk wrote:
>>> I'm new to kepler but love what it can do.
>>>
>>> My question is I would like to use kepler as a web service (WSDL).  
>>> Is  it possible to do that?
>>> For instance I would like to be able to send the service a  
>>> workflow  and input data and have it return the result.
>>>
>>> Stuart Chalk, Ph.D.
>>> Associate Professor of Chemistry
>>> Department of Chemistry and Physics
>>> University of North Florida
>>> 1 UNF Drive
>>> Jacksonville FL 32224
>>> P: 904-620-1938
>>> F: 904-620-1989
>>> E: schalk at unf.edu
>>> W: http://www.unf.edu/coas/chemphys/
>>>
>>> _______________________________________________
>>> Kepler-users mailing list
>>> Kepler-users at ecoinformatics.org
>>> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-users
>>>
>>>
>> _______________________________________________
>> Kepler-users mailing list
>> Kepler-users at ecoinformatics.org
>> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-users
>
> -- 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Matthew B. Jones
> Director of Informatics Research and Development
> National Center for Ecological Analysis and Synthesis (NCEAS)
> UC Santa Barbara
> jones at nceas.ucsb.edu                       Ph: 1-907-523-1960
> http://www.nceas.ucsb.edu/ecoinfo
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~