Transcript Simulation data model

Simulation data model
2007-04-12/13
SNAP data model
Overview
•
•
•
•
•
Goal and approach
Domain model
Logical model
Discussion
Other data models
– IVOA
– Concrete
• Usage in query protocol
2007-04-12/13
SNAP data model
Goals
• Describe spatial simulations for SNAP
( SSSAP ?)
• Use in
– Discovery (simulation registry)
– Query phase in SSSAP protocol
– Meta data in returned data
• See whether we can produce a model
serving other types of simulations as well.
2007-04-12/13
SNAP data model
Modeling approach
• Step 1: Analyse “universe of discourse”,
resulting in a domain model :
– Application independent
– Implementation (language) neutral (UML)
• Step 2: Specialise to requirements, create
logical model
– Application dependent
– Implementation neutral (UML)
• Step 3: create mappings to implementation
languages, physical model:
– XML schema
– Relational database schema
2007-04-12/13
SNAP data model
Simulation domain model
• Based on “domain model for astronomy”
http://www.ivoa.net/internal/IVOA/IvoaDataModel/Domai
nModelv0.9.1.doc
• Main areas/namespaces:
–
–
–
–
–
–
Experiment
Protocol
Phenomenology
Storage
Standards
Model
2007-04-12/13
SNAP data model
Domain model
overview
2007-04-12/13
SNAP data model
Domain model
packages
2007-04-12/13
SNAP data model
Domain model
experiment
2007-04-12/13
SNAP data model
Domain model
protocol
2007-04-12/13
SNAP data model
Domain model
phenomenology
2007-04-12/13
SNAP data model
Experiment:simulation
2007-04-12/13
Protocol::simulation
SNAP data model
Logical model
• Not aim to describe all details of simulations, nor
all types of simulations
• Focus on those aspects of interest to
– discover interesting SNAP services
– query a SNAP service for interesting simulations
– provide meta data for a data product produced by
SNAP service
• How to find out?
Ask scientists (as well as ourselves)!
2007-04-12/13
SNAP data model
Example questions
• What questions would you ask a database
containing (meta-data about) simulations:
–
–
–
–
–
–
–
–
What is simulation about?
What physical processes are implemented?
What values do certain parameters have?
What observables are provided?
What are characteristics of observables?
What type of data products can be returned?
What special data products can be provided as well?
What special services are available on server?
2007-04-12/13
SNAP data model
Logical model
• For our purposes (SNAP), a simulation is
the execution of software that produces a
representation of a spatial system, and
possibly follows its evolution in time by
approximating the true physical processes of
the system with numerical algorithms.
2007-04-12/13
SNAP data model
Logical model
(see wiki page
and doc)
2007-04-12/13
SNAP data model
Discussion
SNAP_DataModel.doc
2007-04-12/13
SNAP data model
Relations to other DMs
• DM WG
– Characterisation
– Quantity (?!)
• Resource Registry WG
– Resource Metadata
• Semantics WG
– UCD
– standard vocabulary
– Ontology of astronomical object types
2007-04-12/13
SNAP data model
Characterisation
• We have a generalisation of (part of?)
characterisation
– a priori and a posteriori
– No support/fill, do we need it? How?
– No accuracy/sensitivity, do we need it? How?
• How do we proceed?
– use Characterisation patterns
– use implementation
• ACTION ITEM
– “tiger team” to complete this aspect of model
2007-04-12/13
SNAP data model
Resource Metadata
• Registration of SNAP resources
• Registration of related services
• Description of simulations published
through SNAP sufficient?
• ACTION ITEM
– “tiger team” for proposing how/what to register
2007-04-12/13
SNAP data model
Semantics
• Various usages of standard vocabularies,
ontologies
• Are the current lists sufficient for
SNAP+Simulation description/querying?
• What is missing?
• ACTION ITEM
– “tiger team” to evaluate this
– propose more concepts
– push for getting such lists accepted in IVOA
2007-04-12/13
SNAP data model
Quantity (?)
• Could this DM play a role in defining how
different data products are stored? Can be
retrieved?
• Possibly suggest standard representations
for complex data structures.
• Possibly suggest bindings to various
popular languages.
• Opinions...?
2007-04-12/13
SNAP data model
Mapping of concrete models
• Need to validate our model by checking whether
concrete models “out there” are properly
supported
–
–
–
–
–
–
–
ITVO data model
Horizon data model
GalICS data model
Data model from OWLS-GIMIC HDF-5 files
GAVO hydrosims data model
Millennium database data model
Example input parameter files
2007-04-12/13
SNAP data model
Usage of data model
• Query protocols
– HTTP
– ADQL
• Serialisation
– XML schema
– Relational database schema
2007-04-12/13
SNAP data model
HTTP GET
• Parameter-value based
• Not simple if underlying model complex (or
is it?)
• At least not in one query  use multiple
queries: drill-down
• Requires more sophisticated (web)
applications, possibly session aware.
• Useful to have this done in a centralised
meta data repository.
2007-04-12/13
SNAP data model
ADQL
• SQL based
• Requires relational mapping (see later)
• Requires supporting a relational database
as well as ADQL
• Again, useful to have a limited number of
meta-data repositories (registry++?).
2007-04-12/13
SNAP data model
Mappings (see doc)
• XML schema
– Mapping prescription
(http://www.ivoa.net/twiki/bin/view/IVOA/VOResource010RevNotes)
– Can be automated (XSLT or ArchitectureWare from XMI
files)
• Relational model
– Simpler and rather many standards for OR mapping
(JDO, Hibernate)
– For ADQL/TAP? need some extra files to provide
database metadata
• If agree on mappings, both can be postponed until
UML is done.
2007-04-12/13
SNAP data model
How to proceed.
2007-04-12/13
SNAP data model
TODO items
• Previously mentioned teams for ironing out certain
issues wrt other IVOA efforts
–
–
–
–
characterisation
ontologies
registry model
quantity
• Test cases
– map concrete models to SNAP model
– implement services
• Gather feedback from community
–
–
–
–
explanatory document
questionnaire
incorporate comments
Euro-VO DCA TEG
2007-04-12/13
SNAP data model