Federation and Fusion of astronomical information Daniel Egret & Françoise Genova, CDS, Strasbourg Standards and tools for the Virtual Observatories.

Federation and Fusion of astronomical information

Daniel Egret &

Françoise Genova,

CDS, Strasbourg

Standards and tools for the Virtual Observatories

Garching VO Conference - June 2002 2

Diversity and Heterogeneity

Specific VO scenarios imply to: cross-match surveys, mission logs,

observational catalogues, personal files collect all pieces of information about

an object or a set of objects build samples of astronomical objects discover rare objects in a

multiwavelength space…

Garching VO Conference - June 2002 3

A specificity of the VO will be to effectively collect data from several diverse and distributed systems:

hence the need for data federation and data fusion.

Garching VO Conference - June 2002 4

Definitions (1): Data Federation

Joining data relevant to the same objects or phenomena,

extracted from archives and databases,

possibly heterogeneous and distributed.

Garching VO Conference - June 2002 5

Definitions (2): Data Fusion

... implies to go one step deeperin the semantic description of the data

so that relevant pieces of information can be immediately compared, merged and/or correlated.

Garching VO Conference - June 2002 6

Outline

We present an overview of current solutions, with examples mainly taken from CDS services and tools : Interoperability tools for data

federation Metadata dictionaries and standards

for data fusion

Garching VO Conference - June 2002 7

Solving a complex query may typically require many steps:

Step 1 : resource discovery : what are the resources that can provide relevant information ?

Step 2 : resource locator : address and query syntax of the resource ?

Step 3 : query processing Step 4 : presentation of the federated

answer (datasets, number of records, pages of information, documentation, ...)

Step 5 : data fusion Step 6 : data visualization.

Garching VO Conference - June 2002 8

1. Resource discovery

On-line archives: SDSS, EIS, 2MASS Object databases : SIMBAD, NED Federated database: VizieR Data centres ; AstroBrowse ; Resource lists: AstroWeb

Garching VO Conference - June 2002 9

Garching VO Conference - June 2002 10

Garching VO Conference - June 2002 11

Generic resource discovery services will be an essential part of the VO.

Garching VO Conference - June 2002 12

2. Resource locator

Locating the most recent (and closest if mirrors) version of archive/survey/table/…

GLU dictionary : description of resource location and query language; astronomy service registry.

Web services (such as Universal Description, Discovery and Integration)... but astronomy specific modules still needed.

Garching VO Conference - June 2002 13

Example ofGLU records describing Aladinservice

Garching VO Conference - June 2002 14

GLU recordsforSkyView

Garching VO Conference - June 2002 15

3. Query processing

Submitting queries to several distributed heterogeneous systems AstroBrowse, AstroGLU, ISAIA Simple Cone search (NVO) Distributed system vs central node ? New protocols : SOAP and WSDL New formats: XML / RPC

Towards distributed (data grid) query processing for the Virtual Observatory.

Garching VO Conference - June 2002 16

Interoperability

A central aspect of the VO is the interoperability : heterogeneous databases and information services do exchange information as part of the query processing.

(see talk by Mark Allen)

Garching VO Conference - June 2002 17

Garching VO Conference - June 2002 18

4. Data presentation

Response from distributed queries: Summary information and dataset

descriptions

Providing multiple responses showing all results in normalized form (e.g. units…) using standard format (FITS, XML) together with documentation files.

Garching VO Conference - June 2002 19

Simbad

Garching VO Conference - June 2002 20

VizieR

Garching VO Conference - June 2002 21

Example of Vizier global search

Garching VO Conference - June 2002 22

5. Data Fusion

Needs a semantic description VOTable format (XML) and UCD (see poster by Sebastien

Derriere)

Example of Aladin tools: overlays, colour composition, astrometric registration and resampling see poster by Fernique et al.

Garching VO Conference - June 2002 23

The ALADIN data integrator

NGC 5236

DSS image HST observation FOV SIMBAD and NED GSC, USNO A2 IUE observations

Garching VO Conference - June 2002 24

GLU system

Garching VO Conference - June 2002 25

Garching VO Conference - June 2002 26

Garching VO Conference - June 2002 27

Aladin:

Chandra

contours

Garching VO Conference - June 2002 28

Aladin:

colour composition

Garching VO Conference - June 2002 29

2MASS combined images

Garching VO Conference - June 2002 30

(see poster by Louys et al.)

Garching VO Conference - June 2002 31

Multi-wavelength cross-identification

Multi-wavelength cross-identifications, at a massive scale, using reference surveys are a key to data fusion.

Interoperable data-mining services.

Garching VO Conference - June 2002 32

6. Data Visualization

Spectral Energy Distribution (NED) Image contours (ALADIN) Sky maps Histograms Colour-magnitude diagrams

Garching VO Conference - June 2002 33

NED: Spectral Energy distribution

Garching VO Conference - June 2002 34

Looking at the Virtual Sky

Garching VO Conference - June 2002 35

Garching VO Conference - June 2002 36

Aladin : All sky

Garching VO Conference - June 2002 37

Garching VO Conference - June 2002 38

Conclusion

At the end of the current VO deployment, we expect VO portals to provide:

resource discovery tools full documentation and library metadata dictionaries normalized query engines.

Data and Information fusion in action…