Webbased technology for storage, processing, and simulation of multi component data in seismology

1
Web-based technology for
storage, processing, and simulation of
multi-component data in seismology

• Robert Barsch
• barsch_at_geophysik.uni-muenchen.de

2
Table of contents

• Motivation
• Project goals
• Working program

3
Motivation
4
State of the Art

• current seismic processing tools (PITSA, Seismic
  Handler, SAC etc.) are mainly limited to classic
  three-component recordings and cannot easily
  handle collocated multi-component data (e.g.,
  pressure, temperature, rotational motions, tilt,
  accelerometer, GPS)
• no standards on how to format, store, distribute
  synthetic seismograms
• computational wave propagation tools for 3D media
  exists (SPICE), but there no tools to provide
  this to observational seismologists and
  interpreters
• currently used seismological databases
• mainly depend on event based data sets (GIANT)
• are badly extensible
• depend on licensed database engines (GIANT,
  Earthworm)
• none of them handle continuous multi-parameter
  data input as well as classical seismic data
• none of them fetch automatically data available
  by data centers or to share a specific data to
  other users

5
Project goals
6
Project goals

• design of a new open source, modular,
  multi-component database structure with access to
  observational infrastructure (easy-extendible,
  future proof, not limited to three-component
  recordings, able to serve event related data
  structures as well as continuous data streams)
• a link between database, multi-component
  processing tools and executable simulation
  algorithms (provided by SPICE)
• development of formats and standards to combine
  the joint storage and processing of observations
  and simulations (partly provided by FDNS and IRIS)

7
Working program
8
Web services

• web services are self-contained and
  self-describing components that can be published,
  discovered, and invoked across the network
• two methods of interaction with WS
• RPC style
• document style (XML document exchange sender may
  not expect or wait for an immediate response)
• independence of platform and implementation
  technology (relying on standards XML, HTTP, SOAP)

9
Web services

• Web Services Definition Language (WSDL)
• provides a complete description of the web
  service, including port, operations, and message
  types involved
• contract between the web service and its
  clients
• allows automated generation of client software
• Simple Object Access Protocol (SOAP)
• standard transport protocol used to invoke a web
  service

10
Web services

• Universal Description, Discovery, and Integration
  (UDDI)
• searchable registry for all available services
• possibility to add service
  (contact data, classifi-
  cation in
  categories)
  and business (license)
  information to
  each
  web service

11
Web services
UDDI
External Apps
TCP/IP
Application Server
DB
WS3
WS2
Disk
WS Client
WS1
12
Web services
UDDI
External Apps
TCP/IP
Application Server
DB
WS3
SOAP
SOAP
WS2
Disk
WS Client
WS1
13
Web services
UDDI
External Apps
TCP/IP
Application Server
DB
WS3
WS2
Disk
WS Client
WS1
14
Web services
WS4
UDDI
External Apps
TCP/IP
Application Server
DB
WS3
WS2
Disk
WS Client
WS1
15
Web Services

• Advantages
• Standards (XML, SOAP, HTTP)
• Developer Tools, i.e. Java Web Services Developer
  Pack (JWSDP) from Sun Microsystems
• Libraries, i.e. implementation of SOAP into 50
  different languages and platforms (2003)
• No firewall problems (HTTP)
• Security authorization, data encryption, HTTPS

16
Web Services

• Disadvantages
• Slower than common solutions
• parsing XML
• data overhead (i.e., SOAP envelop)
• Infrastructure (application server etc.)
• Seismologists ! IT nerds

17
General Development Remarks

• interfaces available to the Internet via web
  services should be strictly defined, so users
  or/and programs can rely on them at any time
• project homepage will host all information to
  connect with these services (e.g., Online
  tutorial, sample clients for different platforms
  and programing languages)
• open source no licensing fees
• source code versioning system (subversion)
• bug tracker forum/wiki

18
Simulation Plug-in

• web-based access to simulation codes (SPICE
  project)
• on request initiate a simulation run on the
  internal supercomputing infrastructure (parallel
  Linux cluster with 150 processors)
• results will be saved into the database and
  therefore publicly available
• user will be informed via a communication gateway
  about the completed simulation run
• only reachable to a certain user group due to the
  valuable processing resource
• uniform, extendible data format for synthetic
  seismograms must be defined

19
Extend Processing Tools

• connect new infrastructure with existing
  processing tools
• SH (user menu)
• Qt-PITSA (cooperation with University Potsdam)
• main focus access and usability of synthetic
  seismograms