Transcript The NVO Compared to Other O`s

THE US NATIONAL VIRTUAL OBSERVATORY
The VO compared to Other O‘s
Jim Gray
Microsoft
NVO Review -- San Diego -- 13-14 Jan 2005
1
The Big Picture: online-science
Experiments &
Instruments
Other Archives
Literature
questions
facts
facts
?
answers
Simulations
•
•
•
•
•
•
The Big Problems
Data ingest
• Query and Visualization tools
Managing a petabyte
• Support/training
Common schema
• Performance
How to organize it?
– Execute queries in a minute
How to reorganize it?
– Batch (big) query scheduling
How to coexist with others?
NVO Review -- San Diego -- 13-14 Jan 2005
2
There Are LOTS of Observatories
•
•
•
•
•
•
•
•
•
•
•
High Energy Physics
Mission To Planet Earth (EOS)
Geology/Seismology
Eventually,
Oceanography
they will all be
Hydrology
integrated into a
Climate
world-wide
Agriculture
digital library
Ecology / Biosphere
Genomics/Proteomics/…
Epidemiology/Medicine / …
And…… Astronomy
NVO Review -- San Diego -- 13-14 Jan 2005
3
Astro VO compared to X VO
Similarities
• Community
Differences
• Community
– Diverse groups do not see common issues
– Building islands of information (stovepipes)
– Need cyber-infrastructure but
• Have no funds to pay for it
• Horrified by price
• Disappointed in past investments
– No charter for data curation & preservation
• Domain
–
–
–
–
–
–
–
Started sooner (80’s)
All literature online
Have data sharing tradition
Smaller (1e4 professionals)
More cohesive
Astronomers know Math & Tech
No commercial/privacy issues
• Domain
– New instruments imply new science methods
– Data is simpler
– Forced to “objectify” the science
– UCD common base vocabulary
– Current tools inadequate for data avalanche
– Survey data volume is huge
(from sensors and simulators)
Systematic and well-documented
– Data avalanche requires
– Good analysis and simulation
• New tools
at least by comparison.
• New algorithms (data mining)
• Data management and lineage (provenance)
• Spatial and temporal data is a central problem.
NVO Review -- San Diego -- 13-14 Jan 2005
4
Personal Observations
• World-Wide Telescope (VO & NVO)
is a model for other domains (widely cited).
• It is important to
–
–
–
–
Document successes and failures
Share tools and techniques
Benefit from other’s advances and tools
Pioneer problem areas
• Common data model / definitions
• Curation
• Preservation
• Tools (visualization, analysis, spatial, temporal, …
NVO Review -- San Diego -- 13-14 Jan 2005
5