VO Data Access Layer

26
VO Data Access Layer IVOA Cambridge, UK 12 May 2003 Doug Tody, NRAO

description

VO Data Access Layer. IVOA Cambridge, UK 12 May 2003. Doug Tody, NRAO. Data access/analysis portal. Link between client (user) data analysis software and the VO Enables distributed multiwavelength data access and analysis Key longer-term goals: Integrate everyday data analysis with VO - PowerPoint PPT Presentation

Transcript of VO Data Access Layer

Page 1: VO Data Access Layer

VO Data Access Layer

IVOACambridge, UK 12 May 2003

Doug Tody, NRAO

Page 2: VO Data Access Layer

Data access/analysis portal

• Link between client (user) data analysis software and the VO

• Enables distributed multiwavelength data access and analysis

• Key longer-term goals:– Integrate everyday data analysis with VO– Achieve a scalable VO architecture (same software

used in multiple contexts)

Page 3: VO Data Access Layer

DAL Scope: Software

• Service data access protocols– Catalog, image, spectrum, etc.

• Service frameworks– Virtual data generation and manipulation, grid computing

• Client side integration– API, Web Services client, components, etc.

• Test cases and demonstrations

Denotes first year priority

Page 4: VO Data Access Layer

DAL Scope: Types of data

• Source catalog• Image• Spectrum, SED• Time series• Event and visibility data

• Key challenges– DAL depends upon registry, query, UCD, DM, representation, etc.– Keep it simple; lower the bar for service implementors

Page 5: VO Data Access Layer

Goals for this meeting

• Agree on DAL scope, WG scope– this will be our kick-off meeting

• Strawman data service architecture– what services are required? (e.g., extend image model, add spectra)

• Define roadmap for next year– drivers, priorities, goals of the partners– define Y1 standards development efforts– test cases and demos

• Example– SIA V1.1+V2, SSA V1, better RG/DM/UCD/QL integration, web

services, test cases

Page 6: VO Data Access Layer
Page 7: VO Data Access Layer

DAL AgendaTime Activity Comments10:00 Intro/Goals Review agenda

10:15 Scope/Priorities/Architecture What standards are needed; common elements, protocols

11:00 BREAK11:30 Spectral Data Access Use cases, priorities; image model vs 1D

spectrum, FITS/XML

13:00 LUNCH14:00 SIA Enhancement Update on Aladin/CDS integration;

priorities

15:00 Roadmap Drivers, priorities

15:30 BREAK16:00 Roadmap Standards efforts, test cases, demos;

standards process

17:00 CLOSE

Page 8: VO Data Access Layer

Goals for this meeting

• Agree on DAL scope, WG scope– this will be our kick-off meeting

• Strawman data service architecture– what services are required? (e.g., extend image model, add spectra)

• Define roadmap for next year– drivers, priorities, goals of the partners– define Y1 standards development efforts– test cases and demos

• Example– SIA V1.1+V2, SSA V1, better RG/DM/UCD/QL integration, web

services, test cases

Page 9: VO Data Access Layer

Existing DAL Prototypes

• Cone search, SIA (NVO)• Jodrell Bank – imaging online from Merlin data

(Anita Richards)• CVO – WFPC assoc, ROSAT, 2QZ (spectroscopic

survey) – generic DA service for all data (query is separate); different service instances, obey same interface

• AUS/VO – HIPASS, ATCA archive, MACHO, SUMSS, 2DF GRSS – Web query i/f

Page 10: VO Data Access Layer

Existing DAL Prototypes

• AVO– SIA, IDHA (all images)

• AstroGrid– HDX (Starlink) – mainly data model, not much data access yet– Also solar data, space physics data

Page 11: VO Data Access Layer

DAL Portal Concept

• Link between client (user) data analysis software and the VO

• Enables distributed multiwavelength data access and analysis

• Key longer-term goals:– Integrate everyday data analysis with VO– Achieve a scalable VO architecture (same software

used in multiple contexts)

Page 12: VO Data Access Layer

DAL Scope: Software

• Service data access protocols– Catalog, image, spectrum, etc.

• Service frameworks– Virtual data generation and manipulation, grid computing

• Client side integration– API, Web Services client, components, etc.

• Test cases and demonstrations

Denotes first year priority

Page 13: VO Data Access Layer

DAL Scope: Types of data

• Dataset (add later)– Query for any type of data in a given ROI– Use other data access services to drill-down to actual data

• Source catalog– May need simplified catalog query in DAL portal– More sophisticated catalog operations elsewhere in VO

• Image– 2D sky images, spectral data cubes, long slit spectra– Access to raw data via VO framework– Data model issues, representation (FITS, Graphic, XML)– Sparsely sampled images including IFU data

• Spectrum, SED• Time series (later?)• Event and visibility data (later?)

• Key challenges– DAL depends upon registry, query, UCD, DM, representation, etc.– Keep it simple; lower the bar for service implementors

Page 14: VO Data Access Layer

DAL Architecture

• Data Access Services– Multiple services

• One service for each view of the data• At a high level may not need to differentiate data

– Integration of common elements of services– Can sometimes view the same data via different services

• E.g., and event list or visibility dataset viewed as an image

– Simple services for most common cases

Page 15: VO Data Access Layer

Application

DAL ClientAdapter

DALService

SimpleServer Grid

ServiceAnalysis

Component

Data MediatorComponent

Data Access Framework

Client Portal

Service Protocols

VO Data Access Layer Architecture

Page 16: VO Data Access Layer

DAL Services

• Key common elements– Registry (caches resource and service data)– Query syntax and capabilities– Metadata– Data models– Data representation (e.g., VOTable)

Page 17: VO Data Access Layer

First cut at priorities for Y1

• SIA V1.1 – Same as now with minor enhancements

• Simple spectral access– Emphasize 1D and SED

• SIA V2.0– General image model (spectral data cubes etc.)– Explore ways to structure metadata (table vs hierarchical)

• Better integration with VO standards– Query standards, registries, MD(UCDs), DM, data representations

• First steps for event, visibility data– E.g., metadata standards to publish raw data

Page 18: VO Data Access Layer

Spectral Data Access

• Goals / Use Cases– discover, retrieve, display spectra for some object– discover, retrieve, display SED for some object– 1D science spectrum for analysis, e.g., classification– multi-wavelength spectral combination– Generate spectra on the fly from grism/radio/HE data– ??

Page 19: VO Data Access Layer

Spectral Data Access

• Priorities– 1D science spectra– SEDs – spectral image cubes (2 spatial + 1 energy)– 2D science spectra (slit spectra)– ??

Page 20: VO Data Access Layer

Spectral Data Access

• Key Issues– general 3/4D image model vs simple 1D/SED– two services versus 1– FITS vs XML (VOTable)

• both should be available for 1D spectra, SEDs (also graphic)• Would need a new FITS standard for 1D spectra to use FITS here

– ??

Page 21: VO Data Access Layer

SIA Evolution

• Interface concepts– roadmap: V1.0, 1.1; V2.0– separation of query/discovery and access– types of services: atlas, pointed, cutout, mosaic– image generation parameters: query ROI supplies default– full specification of IGP possible for mosaic services

Page 22: VO Data Access Layer

SIA Evolution

• Feedback on SIA V1.0– Aladin/CDS integration (Francois)

Page 23: VO Data Access Layer

SIA – Proposed Enhancements

• Registry integration• Image Attributes

– Image provenance and identification - 1– Spectral bandpass (already present) - 1– Spatial bandpass - 3– Resolution - 1– Limiting flux - 2– Image type (future- v2)– Time of observation (1)– Default for case where there are multiple versions of same dataset

• Image Query– Use of image attributes to refine query (e.g., band)– POS, SIZE generalization? Coordinate frameworks

Page 24: VO Data Access Layer

SIA – Proposed Enhancements

• Others– Support image compression in access protocol

Page 25: VO Data Access Layer

Second cut at priorities for Y1

• SIA V1.1 – Same as now with minor enhancements

• Simple spectral access– Emphasize 1D and SED

• SIA V2.0– General image model (spectral data cubes etc.)– Explore ways to structure metadata (table vs hierarchical)

• Better integration with VO standards– Query standards, registries, MD(UCDs), DM, data representations

• First steps for event, visibility data– E.g., metadata standards to publish raw data

• Web services versions of DAL services

Page 26: VO Data Access Layer

Drivers/Priorities

• IAU demo July 03– Add registry support– Add Aladin, other SIA services– Integrate SIA registry into clients (Aladin, DIS, etc.)

• AAS, AVO demos January 04– Web services versions of services?