GGF8 RGIS-RG agenda

Session 1: Data discovery in the grid environment

1. 14:00 -  Welcome
   
2. 14:10  - Scientific Annotation Middleware project, James D. Myers, Pacific Northwest National Laboratory
   
   The Scientific Annotation Middleware (SAM) project ( http://www.scidac.org/SAM/) is developing a layered set of services for managing application data, metadata, pedigree, and annotations on the web/Grid. This talk will highlight several aspects of SAM including its:
   • uniform view of file-based and database content in terms of a typed data resource and associated text/XML key/value pair metadata
   • ability to provide a federated view of multiple databases based on XML descriptions of the mappings
   • ability to dynamically associate new metadata with existing database/file content
   • mechanisms for dynamically configuring internal/external metadata extractors and metadata/data translators
   Examples will be given of how these capabilities are being used in the Collaboratory for Multiscale Chemical Science (CMCS) to provide electronic notebook functionality, portal-based data pedigree browsing capabilities, and lightweight application integration. Related links: http://collaboratory.emsl.pnl.gov/docs/collab/presentations/papers/sam.cise.mayjun03.final.pdf and http://cmcs.ca.sandia.gov
   
   
3. 14-35 - Metadata Management Services for the grid, Stephen Langella, Shannon Hastings, Ohio State University
   
   Over the last few years there has been a great surge of interest in grid computing and the building of 'The Grid'. The grid community describes the grid as a world full of loosely coupled machines running services available to users on the grid. There has been talk in the grid community about metadata protocols that these grid services will adhere to. These metadata protocols are the correct abstraction layer, but who manages these protocols? Who will manage the metadata definitions and metadata that is used in the grid? How will metadata be defined? How will metadata definitions be discovered and who will have the right to distribute, store, and alter them? How will metadata be stored on the grid? Doesn't it seem like the majority, if not all, of data traveling through the grid will be wrapped in metadata? Metadata will enable the data of the grid to be discovered, interpreted, shared, stored, and retrieved in a machine, platform, and protocol independent way. Won't metadata management be needed by almost every service on the grid? Not just for protocol handling but also to represent the data that can flow in and out of these services. It seems like the grid will need a set of services and protocols for which the sole purpose is to manage both the definition of metadata as well as the metadata. This talk will propose a set of services for managing metadata in a grid environment.
   
   
4. 15:00- Transparent Naming and Access to Grid Data Objects, Inderpal Narang, IBM
   
   We propose a data naming and access scheme that provides much greater transparency of not only location and replication, but also name and data format. We first start with a way to reference an arbitrary grid data object and its access mechanism, independent of data type. We then propose a scheme for object discovery where applications contact a single Discovery service with constraints on the object content and the quality of service. This is in contrast to current proposals for a two-step discovery process, where applications contact a catalog service for a logical file handle, and then contact a separate Replica Location Service for a physical file reference. Accompanying paper: Transparent Naming and Access to Grid Data Objects Authors: Vijayshankar Raman, Vitthal Gogate, Inderpal Narang, Mahadevan Subramanian, Shawn Jeffery, Ashmi Chokshi

Session 2: Database use cases and best practices in the grid environment

1. 14:00 - Report on GIS2003 Workshop at NeSC, Edinburgh, Rob Allen, Daresbury and Mark Baker, University of Portsmouth
   
   http://dsg.port.ac.uk/events/workshops/gis2003/
   
   
2. 14:25 - Grid Relational Catalog - GRelC, Sandro Fiore, ISUFI/CACT (Center for Advanced Computational Technologies) University of Lecce, Italy
   
   This is to offer an easy way to manage relational databases in the Globus Community.
                                                                                          
   The GRelC middleware could be very useful for deploying quickly and easily grid-aware applications that interact with Relational Databases.
   
   http://gandalf.unile.it/grelc.htm
   
   
3. 14:50 - General discussion on the milestones and future directions of the group. For example should we include XML as well as Relational models?  Does the charter need revision?