[seek-dev] Meeting at Edinburgh

Matt Jones jones at nceas.ucsb.edu
Tue May 4 13:25:59 PDT 2004


Summary of EcoGrid conference call Tues May 4, 2004

Present: Jones, McCartney, Berkley, Vieglais, Spears, Tao, Zhu

1) EcoGrid agenda items for Scotland meeting
     These supplement the items that Raja sent in his email previously.
     Matt will update the website agenda.

a) resultset structure (see below)
b) GUID (integration with Taxon)
c) globus vs web services -- ridding ourselves of the grid handicap
d) registry metadata -- what's the schema going to be
e) aggregation nodes -- how, where, and why of aggregating content
f) write API & timeline -- how, why of writing to ecogrid nodes
g) EML and EcoGrid -- and other metadata standards interoperation
h) Kepler Ecogrid integration -- overview of progress, critique
i) EcoGrid Point of Presence -- what is an ecogrid node
      (network/computation proximity relative to services)
j) interfaces with SMS

We obviously can't get to all of this in the partial day dedicated to 
EcoGrid, but we can prioritize when we get there.

2) Resultset structure conclusion:
Long discussion about the structure of the resultset schema.  Points of 
agreement include:

1) preserve clear recordId for retrieval from node using get()
2) useful to have addiitonal info culled from document to
    reduce/eliminate get operations for simple displays
3) containment can not be effectively expressed using the simple
    name/value pair system that metacat uses today
4) nested content isn't namespace valid under current system so need
    to revise
5) nested content is v. difficult to do right, but does allow
    containment and order to be expressed
6) returnfield/param approach can be made better by having a
    namespace for the returnfield name, and possibly an xsi:type,
    although the latter is probably not needed if the former exists

Rod will try to formalize his proposed schema in an xsd file and 3 
example result sets: an EML resultset, a Darwin Core resultset, and one 
that mixes records from the two.  That will give us a good basis for 
discussion and decisions in Scotland.

Matt


Arcot Rajasekar wrote:
> This email is mainly for ecogridders but others can peak in :-)
> 
> Hi folks,
> 	We will be meeting in Edinburgh in a litle over a week.
> There is a sessions earlier in the SEEK meeting where we will discuss
> EcoGrid related issues. I think half-a-day is allocated for this.
> 
> What I want for this is some input from our EcoGridders about various
> things:
> 
>     accomplishments
>     demos
>     what is remaining to be done in our previous plan
>     plan to how to proceed beyond what we planned
>     how to integrate with other parts of SEEK (SMS,AMS,Tax,...)
> 
> Please feel free to send ideas and details.
> Each of us will be  giving a demo/talk about what we
> were upto in the last six months  with respect to EcoGrid
> please suggest a topic for your talk.
> 
> 
> Also later in Escience workshop
> we have more Gridding.
>    Intro to EcoGrid   (Matta and Raja)
>    EcoGrid Interfaces (Dave V)
>    Ecogrid Subsystems (Rod,Jing,Bing, someone from Peter's group)
> 
> Following this we have eScince grid folks giving talks on their
> grid projects and after that we have a rump session where we can exchange
> idea about  how we can collaborate and what we can do together.
> 
> Please feel free to discuss these issues also.
> 
> 
> thanks
> 	raja
> 
> 
> To set the ball rolling I will talk about what I have been doing
> which is related or relevant to SEEK:
> 
>    SRB/MCAT development that will provide more functionality
> 	that can be used by SEEK researchers.
>         In particular the work on grid-federation which allows
> 	two autonomous grids (in current case both are SRB grids)
> 	to interact with each other will be of importance as we have
>     	many autonomous systems who want to federate with us.
>      http://www.npaci.edu/dice/Pubs/Zone-SRB.doc
>      http://www.npaci.edu/dice/Pubs/CSI-paper-sent.doc
>    Kepler/Ptolemy - did some experiments with SRB and Ptolemy
>     	mainly with seeing how such things can be integrated.
> 	This can be useful in doing the EcoGrid tp Kepler interfacing
> 	which we will be discussing in detail in Edinburgh
> 
>    Real-time system  data system integartion with SRB. We probably at some
> 	time or other will have to deal with instrument data and
> 	this experience will be useful there.
>      http://www.npaci.edu/dice/Pubs/VORB1.doc
> 
> 
> 
> _______________________________________________
> seek-dev mailing list
> seek-dev at ecoinformatics.org
> http://www.ecoinformatics.org/mailman/listinfo/seek-dev





More information about the Seek-dev mailing list