AstroGrid Group 7: Teemu Toivola Tero Viitala. Problem several separate databases no common interface between databases difficulties of joining related.

AstroGridGroup 7:

Teemu Toivola Tero Viitala

Problem

• several separate databases

• no common interface between databases

• difficulties of joining related data

• limited observatory time resources

• lack of user skills

Shocks seen in the X-ray

Heavy elementsseen in the optical

Dust seen in the IR

Relativistic electrons seen in the radio

Goals

• a working datagrid for key UK databases

• uniform archive query and datamining interface

• the ability to browse simultaneously multiple datasets

• a set of tools for on-line database analysis and exploration

Members• consortium members

– School of Computer Science, Queens University of Belfast (QUB)

– Institute of Astronomy, University of Cambridge

– Institute for Astronomy, University of Edinburgh

– Dept of Physics and Astronomy, University of Leicester

– Mullard Space Science Laboratory, University of London (MSSL)

– Jodrell Bank Observatory, University of Manchester (JBO)

– Space Data Division, Rutherford Appleton Laboratory (RAL)

• financed by PPARC

• executive board AGLI

• day-to-day management

Project content

• Phase A– architecture definition– begin: 1.9.2001– end: 31.12.2002

• Phase B– implementation– begin: 1.1.2003– end: 31.12.2004

Phase A

• WP-A0 Project Management & Infrastructure

• WP-A1 Science Requirements• WP-A2 - A4 Technology assessment• WP-A5 Pilot programs• WP-A6 Support for future databases• WP-A7 Researching existing technologies• WP-A8 Phase B plan• WP-A9 User Interface

Phase B

• Iterations 1 & 2 (June 2003)– basic engine room components– preliminary interface and contents– OGSI experiments

• Iterations 3 & 4 (December 2003)– engine room continued– interface and contents– international program

Phase B• Iteration 5 & 6 (June 2004)

– datamining – workflow– visualisation – OGSA Integration (demo)

• Iteration 7 & 8 (December 2004) – datamining– query Estimator/Optimizer– resource Registry– visualisation– OGSA Integration

Progress

Difficulties

• social– organization– common procedures

• technical– lack of standards– uncertainty of future network bandwidth

• financial– lack of funding

"So Why No Integration Tests? So here we are 2 weeks from code freeze, and still the only bug reports I have on integration testing are from the It03 demo."

"I get the feeling people are doing design -> write code -> reflect -> refactor, document -> test. This is a waterfall model and *not appropriate* (except over a day or two!). We should *not* be designing for next year. We have to get out of the waterfall and into coding!"

Risk P I Comment

Grid software is late or non-functional

H VH

will design for this eventuality so that grid services can be replaced by simple web services