1. Towards a Virtual Observatory Alex Szalay Department of Physics and Astronomy The Johns Hopkins University ADASS 2000

2. The Virtual Observatory National/Global distributed in scope across institutions, agencies and countries available to all astronomers and the public Virtual not tied to a single “brick-and-mortar” location supports astronomical “observations” and discoveries via remote access to digital representations of the sky Observatory general purpose access to large areas of the sky at multiple wavelengths supports a wide range of astronomical explorations enables discovery via new computational tools

3. Why Now ? The past decade has witnessed a thousand-fold increase in computer speed a dramatic decrease in the cost of computing & storage a dramatic increase in access to broadly distributed data large archives at multiple sites and high speed networks significant increases in detector size and performance These form the basis for science of qualitatively different nature

4. Nature of Astronomical Data Imaging –2D map of the sky at multiple wavelengths Derived catalogs –subsequent processing of images –extracting object parameters (400+ per object) Spectroscopic follow-up –spectra: more detailed object properties –clues to physical state and formation history –lead to distances: 3D maps Numerical simulations All inter-related!

5. Trends Future dominated by detector improvements Total area of 3m+ telescopes in the world in m 2, total number of CCD pixels in Megapix, as a function of time. Growth over 25 years is a factor of 30 in glass, 3000 in pixels. Moore’s Law growth in CCD capabilities Gigapixel arrays on the horizon Improvements in computing and storage will track growth in data volume Investment in software is critical, and growing

6. The Age of Mega-Surveys The next generation mega-surveys and archives will change astronomy, due to –top-down design –large sky coverage –sound statistical plans –well controlled systematics The technology to store and access the data is here we are riding Moore’s law Data mining will lead to stunning new discoveries Integrating these archives is for the whole community => Virtual Observatory

7. Ongoing surveys Large number of new surveys –multi-TB in size, 100 million objects or more –individual archives planned, or under way Multi-wavelength view of the sky –more than 13 wavelength coverage in 5 years Impressive early discoveries –finding exotic objects by unusual colors L,T dwarfs, high-z quasars –finding objects by time variability gravitational microlensing MACHO 2MASS DENIS SDSS GALEX FIRST DPOSS GSC-II COBE MAP NVSS FIRST ROSAT OGLE... MACHO 2MASS DENIS SDSS GALEX FIRST DPOSS GSC-II COBE MAP NVSS FIRST ROSAT OGLE...

8. The Discovery Process discover significant patterns from the analysis of statistically rich and unbiased image/catalog databases understand complex astrophysical systems via confrontation between data and large numerical simulations Past: observations of small, carefully selected samples of objects in a narrow wavelength band Future: high quality, homogeneous multi-wavelength data on millions of objects, allowing us to The discovery process will rely heavily on advanced visualization and statistical analysis tools

9. The Necessity of the VO Enormous scientific interest in the survey data The environment to exploit these huge sky surveys does not exist today! –1 Terabyte at 10 Mbyte/s takes 1 day –Hundreds of intensive queries and thousands of casual queries per-day –Data will reside at multiple locations, in many different formats –Existing analysis tools do not scale to Terabyte data sets Acute need in a few years, solution will not just happen

10. VO- The challenges Size of the archived data 40,000 square degrees is 2 Trillion pixels –One band 4 Terabytes –Multi-wavelength 10-100 Terabytes –Time dimension10 Petabytes Current techniques inadequate –new archival methods –new analysis tools –new standards Hardware/networking requirements –scalable solutions required Transition to the new astronomy

11. VO: A New Initiative Priority in the Astronomy and Astrophysics Survey Enable new science not previously possible Maximize impact of large current and future efforts Create the necessary new standards Develop the software tools needed Ensure that the community has network and hardware resources to carry out the science Keep up with evolving technology

12. New Astronomy- Different! Data “Avalanche” –the flood of Terabytes of data is already happening, whether we like it or not –our present techniques of handling these data do not scale well with data volume Systematic data exploration –will have a central role –statistical analysis of the “typical” objects –automated search for the “rare” events Digital archives of the sky –will be the main access to data –hundreds to thousands of queries per day

13. Examples: Data Pipelines

14. Examples: Rare Events Discovery of several new objects by SDSS & 2MASS SDSS T-dwarf (June 1999)

15. Examples: Reprocessing Gravitational lensing 28,000 foreground galaxies over 2,045,000 background galaxies in test data (McKay etal 1999)

16. Examples: Galaxy Clustering Shape of fluctuation spectrum –cosmological parameters and initial conditions The new surveys (SDSS) are the first when logN~30 Starts with a query Compute correlation function –All pairwise distances N 2, N log N possible Power spectrum –Optimal: the Karhunen-Loeve transform –Signal-to-noise eigenmodes –N 3 in the number of pixels Likelihood analysis in 30 dimensions: needs to be done many times over

17. Relation to the HEP Problem Similarities –need to handle large amounts of data –data is located at multiple sites –data should be highly clustered –substantial amounts of custom reprocessing –need for a hierarchical organization of resources –scalable solutions required Differences of Astro from HEP –data migration is in opposite direction –the role of small queries is more important –relations between separate data sets (same sky) –data size currently smaller, we can keep it all on disk

18. Data Migration Path portal HEP Astro Tier 0 Tier 1 Tier 2 Tier 3

19. Queries are I/O limited In our applications few fixed access patterns –one cannot build indices for all possible queries –worst case scenario is linear scan of the whole table Increasingly large differences between –Random access controlled by seek time (5-10ms), <1000 random I/O /sec –Sequential I/O dramatic improvements, 100 MB/sec per SCSI channel easy reached 215 MB/sec on a single 2-way Dell server Often much faster to scan than to seek Good layout => more sequential I/O

20. Fast Query Execution Query plan –given the layout of the database, how can one get the result the fastest possible way –evaluate a cost function, using a heuristic algorithm Indexing –clustered and unclustered database layout is crucial –one-dimensional indices: B-tree, R-tree –higher dimensional indices: Quad-tree, Oc-tree,KD-tree, R+-tree –typically logN access instead of N

21. Distributed Archives Network speeds are much slower –minimize data transfer –run queries locally I/O will scale almost linearly with nodes –1 GB/sec aggregate I/O engine can be built for <$100K Non-trivial problems in –load balancing –query parallelization –queries across inhomogeneous data sources These problems are not specific to astronomy –commercial solutions are around the corner

22. User Interface Analysis Engine Master Objectivity RAID Slave Objectivity RAID Slave Objectivity RAID Slave Objectivity RAID Slave SX Engine Objectivity Federation SDSS: Distributed Layout

23. Astronomy Issues What are the user profiles How will the data be analyzed How to organize searches Geometric indexing Color searches and localization SDSS Architecture Virtual data

24. User Profiles Power users –sophisticated, with lots of resources –research is centered around the archival data moderate number of very large queries, large output sizes Astronomers –frequent, but casual lookup of objects/regions –archives help their research, but not central large number of small queries, cross-identification requests Wide public –browsing a `Virtual Telescope’ –can have large public appeal –need special packaging very large number of simple requests

25. Geometric Approach Main problem –fast, indexed searches of Terabytes in N-dim space –searches are not axis-parallel simple B-tree indexing does not work Geometric approach –use the geometric nature of the data –quantize data into containers of `friends’ objects of similar colors close on the sky clustered together on disk –containers represent coarse-grained map of the data multidimensional index-tree (eg KD-tree)

26. Geometric Indexing “Divide and Conquer” Partitioning 3  N  M3  N  M 3  N  M3  N  M Hierarchical Triangular Mesh Split as k-d tree Stored as r-tree of bounding boxes Using regular indexing techniques AttributesNumber Sky Position 3 Multiband FluxesN = 5+ Other M= 100+ AttributesNumber Sky Position 3 Multiband FluxesN = 5+ Other M= 100+

27. Computing Virtual Data Analyze large output volumes next to the database –send results only (`Virtual Data’): the system `knows’ how to compute the result (Analysis Engine) Analysis: different CPU to I/O ratio than database –multilayered approach Highly scalable architecture required –distributed configuration – scalable to data grids Multiply redundant network paths between data-nodes and compute-nodes –`Data-wolf’ cluster => Virtual Data Grid

28. A Data Grid Node Hardware requirements Large distributed database engines –with few Gbyte/s aggregate I/O speed High speed (>10 Gbit/s) backbones –cross-connecting the major archives Scalable computing environment –with hundreds of CPUs for analysis Objectivity RAID Objectivity RAID Objectivity RAID Objectivity RAID Objectivity RAID Objectivity RAID Objectivity RAID Database layer 2 GBytes/sec Compute node Compute layer 200 CPUs Interconnect layer 1 Gbits/sec/node Other nodes 10 Gbits/s

29. GriPhyN GriPhyN - Grid for Physics Networks Goal: build a prototype Virtual Data Grid Create ~20 Tier 2 centers for the data processing Includes ATLAS, ALICE, CMS, LIGO and SDSS Principals –Paul Avery, Ian Foster, Harvey Newman, Carl Kasselman –collaboration of 16 universities and laboratories Phase 1 funding announced a week ago –NSF ITR Program: $11.9M Total scope about $60M www.griphyn.org

30. Clustering of Galaxies Generic features of galaxy clustering: Self organized clustering driven by long range forces These lead to clustering on all scales Clustering hierarchy: distribution of galaxy counts is approximately lognormal Scenarios: ‘top-down’ vs ‘bottom-up’

31. Clustering of Computers Problem sizes have lognormal distribution –multiplicative process Optimal queuing strategy –run smallest job in queue –median scale set by local resources: largest jobs never finish Always need more computing –‘infall’ to larger clusters nearby –asymptotically long-tailed distribution of compute power Short range forces: supercomputers Long range forces: onset of high speed networking Self-organized clustering of computing resources –the Computational Grid

32. VO: Conceptual Architecture Data Archives Analysis tools Discovery tools User Gateway

33. The Flavor/Role of the NVO Highly Distributed and Decentralized Multiple Phases, built on top of another Establish standards, meta-data formats Integrate main catalogs Develop initial querying tools Develop collaboration requirements, establish procedure to import new catalogs Develop distributed analysis environment Develop advanced visualization tools Develop advanced querying tools

34. NVO Development Functions Software development –query generation/optimization, software agents, user interfaces, discovery tools, visualization tools Standards development –Meta-data, meta-services, streaming formats, object relationships, object attributes,... Infrastructure development –archival storage systems, query engines, compute servers, high speed connections of main centers Train the Next Generation –train scientists equally at home in astronomy and modern computer science, statistics, visualization

35. The Mission of the VO The Virtual Observatory should  provide seamless integration of the digitally represented multi-wavelength sky  enable efficient simultaneous access to multi- Terabyte to Petabyte databases  develop and maintain tools to find patterns and discoveries contained within the large databases  develop and maintain tools to confront data with sophisticated numerical simulations

36. Sociological Impact The VO is an entirely new way of doing astronomy Will have a serious sociological impact One can expect a love-hate response from the community Needs to be driven by science goals Technology will constantly evolve Educational impact two-fold –need new skills to use it efficiently –provides enormous new opportunities Astronomy is rather unique in its wide appeal –the VO can reach out to an even wider audience

37. Conclusions Databases became an essential part of astronomy: most data access will soon be via digital archives Data at separate locations, distributed worldwide, evolving in time: move queries not data (if you can)! Computations in both processing and analysis will be substantial: need to create a `Virtual Data Grid’ Problems similar to HEP, lot of commonalities, but data flow more complex Interoperability of archives is essential: the Virtual Observatory is inevitable www.voforum.orgwww.voforum.org