Component Definition
Code	AgCd10	Name	Data Federation

Forecast Start	01-Jul-2003	Actual Start	01-Jul-2003
Forecast End	30-Sep-2003	Actual End	dd-mmm-yyyy

Description

A discussion paper on federation and what it means in practice as applied to astronomical datasets is contained in DataFederation. The proposed work divides naturally into two sections

• How to cross-match source catalogues using a spatial join
• How to unify or merge the results of querying diverse and distributed resources

The former will be the main responsibility of Clive Page, the latter will be mainly carried out by Patricio Ortiz, but considerable feedback in both directions is anticipated.

Because this is a research project it is less easy to determine in advance exactly how much time needs to be devoted to each sub-section, but our initial aim is to proceed as in the following time-table:

Week	Dates	Patricio Ortiz	Clive Page
1	July 1-4	Analysis of problem	Analysis of problem
2	July 7-11	Plan work	Plan work
3	July 14-18	Set up PC and cluster to send/retrieve data using Datoz as DBMS	Select trial datasets and convert to TSV format
4	July 21-25	Populate databases, adapt Datoz to produce required output	Test Postgres joins using R-tree and multi-column index
5	July 28-Aug 1	(on holiday)	Continue Postgres tests using HTM/HEALpix mapping functions
6	Aug 4-8	(on holiday)	Test Postgres stored procedures for join functionality
7	Aug 11-15	(on holiday)	Test Mysql joins using R-tree and multi-column index
8	Aug 18-22	Get functional queries and data retrieval	Continue Mysql tests using mapping functions
9	Aug 25-29	Design user interface for multi-catalogue cone-search etc	Test DB2 joins using selected methods
10	Sep 1-5	Work on identifying equivalent quantities	Continue testing DB2
11	Sep 8-12	continuation of above	Write up and publish cross-matching results
12	Sep 15-19	Modify GUI to allow user choice on automatically generated column matches	Design generic cross-matching facility
13	Sep 22-30	completion of above work: write up and publish results	Start design of generic cross-match facility

Deliverables and Milestones
Code	D/M	Name	Status	Due
AgCd10-001	DataFederationandDataMining Report on data federation and data mining	CGP	complete	26-Sep-2003
AgCd10-002	DataFedeReport Report on federation of diverse datasets	PFO	complete	30-Sep-2003