The task of being able to pass data by reference to a webservice is not something I'm currently working on for SARA/NBIC, although it was indicated as being one of the options for future work. Currently I am working on setting up webservices that run jobs on Globus grids, and then writing a Taverna plugin to manage the certificate delegation required to access such a service.
I'm already aware of how to fix the data reference problem for local activities (i.e. not webservices) using the new T2 Platform, as it is fairly easy to write new kinds of data reference handlers for it which can intercept the various file operations, but this is no good for the T2 GUI which does not yet use the T2 Platform (and so plugins are incompatible between the two), and neither does it solve the webservices issue (which would require both a new T2 GUI plugin and new kinds of webservices, just as I've been doing to solve the Globus issue). Also I'm not sure if this would fix existing activities as (I think) each activity's source code would still need to explicitly specify in it's list of acceptable input port reference types if it wants to be able to natively handle files that are passed in by reference. I'm visiting Machiel and Pieter at SARA on April 15th-17th during which we'll discuss what I should be doing for them during May-July. Anything in particular you'd like me to do, just ask them and they'll try to work it into their list of priorities for me to work on. cheers, Richard Marco Roos wrote: > Hello Spiros and Richard, > > This seems like the kind of problem you are addressing, isn't it? Would > your solution be a solution for Ulf? > If so, could you explain a little bit how you would address his problem? > > Many thanks! > Marco. > > > Paolo Missier wrote: >> Hi, >> me again, reporting on a nice and friendly chat with Ulf Leser (Uni. >> Humboldt, long experience in the DILS area in addition to straight data >> management research), who among other things described his difficulties >> in dealing with large data transfers in and out of the Taverna (1.7) >> space, while executing workflows that encode microarray pipelines. The >> problem is that large data files are repeatedly pulled from a server >> into the Taverna space, used by a service and then the results are >> pushed back to the server again, instead of having Taverna pass data >> files server references to the services that need them. This is because >> the services are not natively aware of the data location. Their solution >> so far has been to wrap them using Soaplab. While service wrappers seems >> like the only viable solution to me, they were not particularly happy >> with it. Any smarter suggestion on this from the group? Couldn't >> perhaps such wrappers, which encode a standard pattern, be >> auto-generated from service and data annotations? >> >> -Paolo > -- Richard Holland, BSc MBCS Finance Director, Eagle Genomics Ltd T: +44 (0)1223 654481 ext 3 | E: [email protected] http://www.eaglegenomics.com/ ------------------------------------------------------------------------------ _______________________________________________ taverna-hackers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/taverna-hackers Developers Guide: http://www.mygrid.org.uk/usermanual1.7/dev_guide.html FAQ: http://www.mygrid.org.uk/wiki/Mygrid/TavernaFaq
