WP-A2: Data Grid Technology

Objectives

• Identify key component systems and technologies.
• Assess these components via prototyping and test-bed programs.
• Contribute the lessons learned to the definition of Phase-B.

Inputs

• Use cases from WP-A1.
• Information from other Grid projects.
• A range of existing prototypes ranging from those using CGI and HTTP (e.g. the set that work with GAIA), to those using more specialised protocols.
• Data-sets hosted at the sites of the members of the AstroGrid consortium, and in particular those from WP-A5, around which can be built new data-services using the trial technologies.
• Grid-related products from outside AstroGrid ranging from high-level toolkits such as GridPort^?, to lower level ones such as Globus, SRB, and underlying technologies such as JINI, SOAP, CORBA. In addition there are operational systems such as Vizier from CDS.

Tasks

• We will evaluate the work being done by in a number of advanced projects such as GridPP and DataGrid, work undertaken by Computer Science Colleagues, and related standards produced by organizations such as w3C, e.g. RDF.
• We expect these evaluations will reveal key components that work independently of the scientific application and can be shared between AstroGrid and other e-science projects, both data-grids and compute-grids. Shared components are likely to cover authentication, resource allocation, scheduling, transport issues, generic management of distributed tasks and generic data-discovery tools.
• We recognise that some or all of these components may need to be specific to astronomy. Particular attention needs to be paid to the data-models for transport of data and metadata, possibly using FITS and XML, and to the way in which queries are phrased, which needs to deal with astronomical measures of position and spectral intensity. We therefore expect to contribute a limited number of new or enhanced components to the field of grid technology.
• Each component that shows sufficient promise to be considered for AstroGrid Phase-B will be evaluated in one of a limited series of test-beds, and each test-bed should be developed to the point where it can be used in a beta test with scientific users.
• There will be interactions with other work packages; e.g. the definition of a distributed query mechanism may depend upon the database systems used in WP-A4.

Outputs

• Reports on the evaluation of component technologies.
• For each of the test-beds, a report on its strengths and weaknesses, a description of end-user reactions, and a projection of the likely results if the system were to be scaled up first to national and then to international operation.
• Recommendations for standard metadata to be adopted in AstroGrid Phase-B. This should be a joint report with WP-A4 and WP-A5.
• A report on how the existing use-cases were satisfied by the various test-beds.
• A list of additional use-cases arising from the trials, particular from the beta-tests.
• A report on how our recommended solution could benefit other projects building data-grids.