Requirements from astronomy in the Virtual Observatory era

1
Requirements from astronomy in the Virtual
Observatory era

• Bob Mann
• Institute for Astronomy NeSC
• University of Edinburgh

2
Outline of talk

• What is the Virtual Observatory (VObs)?
• How is it changing astronomy?
• Who in astronomy needs e-Science skills?
• Data centre staff
• VObs middleware developers
• Data analysis tool developers
• Research astronomers
• Summary
• What skills are needed?
• How can these requirements be met?

3
What is the Virtual Observatory?

• A goal
• An interoperable federation of the worlds
  astronomical data sources
• A standards agency/coordination body
• International Virtual Observatory Alliance
• Its members are VObs projects in Australia,
  Canada, China, France, Germany, Hungary, India,
  Italy, Japan, Korea, Russia, Spain, UK, USA, and
  the EU.

4
Why is the VObs needed?

• We observe across the whole electromagnetic
  spectrum
• Different views of a local spiral galaxy
• Need all of them to understand its physics fully
• The databases holding them are located all over
  the world

5
Schematic of the VObs
DB 4
DB 3
DB 2

DB 5
DB 1
Compute Resource 1
App 1
Registry
App 3
App 2
Workflow
Compute Resource 2
App 4
User
Portal
App 5
6
What services will the Virtual Observatory need?

• Data discovery
• Looking up metadata in a registry
• Data access
• With/without authentication authorization
• Data integration
• Cross-matching entries in databases
• Data manipulation
• Data mining, data analysis, etc

7
Data centre staff
DB 4
DB 3
DB 2

DB 5
DB 1
Compute Resource 1
App 1

• Data Curation
• Provenance, DBMS design and operations
• Data Access
• Security, Registry Metadata, Web Services
• Data Manipulation
• User-uploaded code, job scheduling

Registry
8
VObs middleware developers
Web/Grid services for the discovery, transport
and integration of data Workflow service
composition Metadata and service registration
Job management Security
DB 2

DB 1
Registry
App 3
Workflow
Compute Resource 2
Portal
9
Data analysis tool developers

• Data Access
• OGSA-DAI/data access web services
• Service Registration
• Integration into workflows
• Job submission

DB 1
Registry
App 3
Workflow
Compute Resource 2
10
Research astronomers

• Initially
• Use of portal
• Soon
• Script and tool development using APIs, local
  job control, wrapping apps as web services

DB 2

DB 1
Registry
Workflow
Compute Resource 2
App 4
User
Portal
App 5
11
Summary of skills needed

• Web/Grid services for data discovery, access,
  integration and manipulation
• Metadata for data service registration
• Data curation provenance, preservation
• Workflow creation and execution
• Job control local and remote
• Security

12
How can these needs be met? (1)

• Learning through experience
• Difficult fast moving field where to start?
• Danger of learning the wrong technologies
• Inefficient undirected training can be slow
• Hiring people with appropriate skills
• Not many people with e-Science skills yet
• and expensive to employ on research grants
• Really need astronomers for some jobs
• Danger of losing scientific direction to projects

13
How can these needs be met? (2)

• Existing staff training courses
• Can be expensive and time-consuming
• Need to budget time and money for them
• Funders need to understand their necessity
• Future staff education
• Astronomers MSc-level courses either as MSc or
  as part of PhD training
• IT specialists inclusion of e-science content in
  CS degree programmes