German Astrophysical Virtual Observatory Overview and Results So Far W. Voges, G. Lemson, H.-M. Adorf
USM 13-Feb-2004 Now: N x M interfaces
USM 13-Feb-2004 Future: N + M interfaces
USM 13-Feb-2004 (GA)VO architecture
USM 13-Feb-2004 International Virtual Observatory Alliance
USM 13-Feb-2004 IVOA working groups Registry Data Access Layer Data Modeling VOQL/ADQL VOTable UCD Grid and Web services
USM 13-Feb-2004 Standards and Protocols IVOA exchange standards Simple Cone Search (SCS) Simple Image Access (SIA) Simple Spectra Access (SSA) VOTable OpenSkyNode, VOQL/ADQL General standards HTTP, XML, SOAP, WSDL
USM 13-Feb-2004 Registry
USM 13-Feb-2004 SCS
USM 13-Feb-2004 Result
USM 13-Feb-2004 VOTable
USM 13-Feb-2004 Tools – Aladin
USM 13-Feb-2004 Data Model What? Why? Observation Quantity Simulation Domain model/ontology as “Esparanto” Archive meta-data repositories
USM 13-Feb-2004 GAVO’s Pillars Federation of datasets Matching & Classification & Visualization (NGSE) Theoretical VO and Theory-Observation Interface Grid computing
USM 13-Feb-2004 Federation of Datasets Towards a multi-wavelength view of universe Matching Source extraction (P. Schuecker et al.) Classification (GAVO/ClassX) SED’s ADQL/OpenSky Node
USM 13-Feb-2004 Federation of Datasets (cntd.) Web services Allow J2EE and.NET applications to interact with each other Based on standards + protocols HTTP + XML or SOAP + WSDL Examples MCMCS download manager
USM 13-Feb-2004 NGSE (Next Generation Search Engine) Download Manager Retrieves data from multiple distributed databases Matcher Matches sources based on sky-position (astronomical sources have no unique identifier) Classifier Uses multi-wavelength data for identification purposes Visualization Tool
USM 13-Feb-2004 Theory in the VO Publication of simulations Exposing theory services Theory-observational interface “Virtual telescope”, e.g. for proposal preparation support
USM 13-Feb-2004 Grid computing Working with Globus Toolkit Looking into alternative technologies SGE, EJBs Gadget simulation running on joint MPE and AIP workstation clusters Interested in a good application that would benefit from parallel distributed computing: too large to run on a single computer, or require a quick turn around
USM 13-Feb-2004 GAVO Accomplishments Publication of ROSAT datasets Download Manager Matcher Documents ADASS contributions TVO White Paper WIP VOTable readers and writers Service “adapters”
USM 13-Feb-2004 ROSAT BSC, FSC & Photons VO-compatibility: Simple cone search in VOTable out New features Exposure times Galactic & Cartesian coordinates Healpix indices
USM 13-Feb-2004 MCMCS Download Manager Tool … … accesses registry at JHU User … … selects up to 70 distributed catalogues … specifies one or more sky-locations Tool … … queries remote catalogues … retrieves datasets (VOTables) for further processing
USM 13-Feb-2004 MCMCS Download Manager
USM 13-Feb-2004 MCMCS Download Manager
USM 13-Feb-2004 Workflow
USM 13-Feb-2004 Matcher (on-going work) Essential for data mining Prototype features Performs “fuzzy” match between pairs of source lists from different catalogues Computes probability of real match based on astrometric uncertainties Moving matcher into production use Currently exported to Canadian Virtual Observatory Feeding matches to ClassX classifier Long-term goal: assemble an SED
USM 13-Feb-2004 Matcher (on-going work)
USM 13-Feb-2004 Classification (soon) ClassX: international collaboration Requires data from several large sky- surveys X-ray: ROSAT (BSC + FSC) Optical: SDSS, USNO Infrared: 2MASS Radio: FIRST, NVSS, SUMSS True multi-wavelength VO-application
USM 13-Feb-2004 Theory in the VO TVO white paper Formation of an IVOA special interest group
USM 13-Feb-2004 Outlook Emphasis on 3 rd generation REST- or SOAP-based Web services Proxy/adapter services, e.g. to SDSS, CDS, NED, … “Virtual telescope” for proposal preparation support 4 th generation “Grid” application: combination of several Web services into one, e.g. Upload of coordinate list to major datacenters Combine match lists using “fuzzy” matcher Construct quasi SED Feed match candidates into remote ClassX classifier Plan: offer as public service
USM 13-Feb-2004 Collaboration Models Listen to scientists/end user requests Support data publication Implement new services E.g. decision support tool for (observing or funding) proposal preparation
USM 13-Feb-2004 How could we help? Knowledge transfer Publishing datasets in a VO-compatible form Establishing new services (e.g. wrappers around legacy code) Offer expertise with Web and J2EE apps, protocols and new standards
USM 13-Feb-2004 Questions What do you expect from us? What role do you see yourself in? Scientist/end user or data provider? What could you offer? How can we collaborate?
USM 13-Feb-2004 Discussion