Workflow Enactment of Grid-Enabled Geospatial Web Services

1
Workflow Enactment of Grid-Enabled Geospatial Web
Services

• Gobe Hobona, David Fairbairn, Philip James

UK e-Science All Hands Meeting - 12th September
2007 - Nottingham
2
Overview

• Background
• OGC Web Services
• Workflow Enactment
• SAW-GEO Architecture
• User interface Geoportal
• Conclusions

3
Background

• Open Geospatial Consortium (OGC)
• ISO19119 and OGSA
• JISC Grid/OGC Collision Programme
• Security SEE-GEO
• Workflow SAW-GEO
• OGC Interoperability Experiments (OWS-4)

4
OGC Web Services

• Web Feature Services (WFS)
• Disseminates vector geospatial data
• Web Map Services (WMS)
• Generates geovisualisations/maps from any
  geo-data source
• Web Coverage Services (WCS)
• Disseminates raster geospatial data
• Web Processing Services (WPS)
• Runs geocomputational models or geospatial
  operations on user-supplied datasets

5
Sample OWS Request Messages
ltGetCoverage service"WCS" version"1.0.0"
xmlns"http//www.opengis.net/wcs"
xmlnsnurc"http//www.nurc.nato.int"
xmlnsogc"http//www.opengis.net/ogc"
xmlnsgml"http//www.opengis.net/gml"
xmlnsxsi"http//www.w3.org/2001/XMLSchema-instan
ce" xsischemaLocation"http//www.opengis.net/
wcs http//schemas.opengis.net/wcs/1.0.0/getCovera
ge.xsd"gt ltsourceCoveragegtnurcArc_Samplelt/sourceCo
veragegt ltdomainSubsetgt ltspatialSubsetgt ltgmlEnvelo
pe srsName"EPSG4326"gt ltgmlposgt9.420000076293944
42.20000076293945lt/gmlposgt ltgmlposgt11.819999694
824217 43.90000049273173lt/gmlposgt lt/gmlEnvelopegt
.. lt/domainSubsetgt ltrangeSubsetgtltaxisSubset
name"Band"gtltsingleValuegt1lt/singleValuegtlt/axisSubs
etgtlt/rangeSubsetgt ltoutputgt ltcrsgtEPSG4326lt/crsgtltre
sponseCrsgtEPSG32626lt/responseCrsgtltformatgtTIFFlt/fo
rmatgt lt/outputgt lt/GetCoveragegt
ltwfsGetFeature service"WFS" version"1.0.0"
outputFormat"GML2" xmlnstopp"http//www.openp
lans.org/topp" xmlnswfs"http//www.opengis.net
/wfs" xmlnsogc"http//www.opengis.net/ogc"
xmlnsgml"http//www.opengis.net/gml"
xmlnsxsi"http//www.w3.org/2001/XMLSchema-instan
ce" xsischemaLocation"http//www.opengis.net/w
fs http//schemas.opengis.ne
t/wfs/1.0.0/WFS-basic.xsd"gt ltwfsQuery
typeName"toppstates"gt ltwfsPropertyNamegttopp
STATE_NAMElt/wfsPropertyNamegt
ltwfsPropertyNamegttoppPERSONSlt/wfsPropertyNamegt
ltogcFiltergt ltogcBBOXgt
ltogcPropertyNamegtthe_geomlt/ogcPropertyNamegt
ltgmlBox srsName"http//www.opengis.net/gml/s
rs/epsg.xml4326"gt ltgmlcoordinatesgt-75
.102613,40.212597 -72.361859,41.512517lt/gmlcoordi
natesgt lt/gmlBoxgt lt/ogcBBOXgt
lt/ogcFiltergt lt/wfsQuerygt lt/wfsGetFeaturegt
6
Sample OWS Response Messages
lt/gmlfeatureMembergt ltgmlfeatureMembergt lttopps
tates fid"states.40"gt lttoppthe_geomgt ltgmlMultiP
olygon srsName"http//www.opengis.net/gml/srs/eps
g.xml4326"gt ltgmlpolygonMembergt ltgmlPolygongt ltgm
louterBoundaryIsgt ltgmlLinearRinggt ltgmlcoordinat
es decimal"." cs"," ts" "gt -77.476082,39.719528
-78.096222,39.725368 74.984978,41.099285
-74.989265,41.081665 -74.966675,41.082581
-77.476082,39.719528 lt/gmlcoordinatesgt lt/gmlLine
arRinggt lt/gmlouterBoundaryIsgt lt/gmlPolygongt lt/gm
lpolygonMembergt lt/gmlMultiPolygongt lt/toppthe_ge
omgt lttoppSTATE_NAMEgtPennsylvanialt/toppSTATE_NAME
gt lttoppPERSONSgt1.1881643E7lt/toppPERSONSgt lt/topp
statesgt lt/gmlfeatureMembergt ltgmlfeatureMembergt
.

• NCOLS 522
• NROWS 300
• XLLCENTER -1.6878262291489
• YLLCENTER 54.93369175561999
• CELLSIZE 3.013533976228086E-4
• NODATA_VALUE -9999.0
• 62.20000 61.70000 61.20000 61.20006 60.7000 60.400
  00 60.40000 60.09999 59.5
• 61.79999 61.5 61.09999 61.09999 60.7000 60.40000 6
  0.40000 60.20000 60.09999847
• 61.79999 61.5 61.099998 61.09999 60.7999 60.5 60.5
  60.5 60.5
• 61.90000 61.59999 61.20000 61.20000 60.9000 60.700
  00 60.70000 60.79999 60.5
• 62.5 62.20000 61.799999 61.79999 61.79999 61.4000
  0 61.40000 61.09999 60.90000153
• 62.70000 62.40000 62.40000 62.40000 62 61.70000 61
  .70000 61.40000 61.09999847
• 63.09999 63 62.59999 62.59999 62.4000 62 62 61.799
  99 61.5
• 63.70000 63.5 63.20000 63.20000 62.9000 62.59999 6
  2.59999 62.40000 62.09999847
• 63.90000 63.70000 63.40000 63.400001 63.0999 62.90
  000 62.90000 62.59999 62.40000153
• 64.19999 64 63.70000 63.70000 63.5 63.29999 63.299
  999 63.09999 62.79999924
• 64.40000 64.19999 64 64 63.7000 63.59999 63.599997
  63.29999 63.09999847
• 64.90000 64.69999 64.5 64.5 64.4000 64.19999 64.19
  999 64 63.79999924
• 64 64.90000 64.80000 64.80000 64.5999 64.5 64.5 6
  4.40000 64.19999695

7
Grid-enabling OWS

• SOAP-based proxy service
• Created from OGC XML schemas (XSD)
• Requests create temporary datasets
• Resource management at two levels
• Temporary datasets managed by proxy service
• Persistent data managed through OWS e.g. WFS-T
• Security limitations
• OWS currently do not offer message encryption
• Approach to wrap OWS in firewall and allow
  external access only through proxy service

8
Geospatial Data is Very Large

• Vector data
• Thousands of coordinates
• Multiple dimensions
• Raster data
• Thousands of pixels (very high resolution)
• Multispectral imagery (200 band imagery)
• Solution
• Pass data by URL references and get client to
  stream data in

9
Load Experiment
10
Workflow Enactment

• Recognised by both ISO19119 and OGSA
• Several options for workflow enactment
• SCUFL, BPEL, Keppler etc
• Selected BPEL because
• OASIS Standard, i.e. WS-BPEL 2.0
• Multi-vendor support including IBM, Sun
  Microsystems, ActiveEndpoints, Oracle etc
• Availability of open source enactors

11
WS-BPEL Overview

• Based on Activities
• Basic or Structured
• Variables
• Declaration
• Assignment
• Conditionals
• Loops
• Fault handlers

12
An Example Geospatial Workflow
WPS 1 Generalise
WFS
Workflow Enactor
client
WPS 2 Clip
Based on OGC OWS-4 GeoProcessing Workflow Scenario
13
Possible Applications for Geo-Workflows

• Emergency Management
• Where each activity depends on the result of a
  previous activity
• Geographic modelling
• Where several steps are needed before a final
  model is produced e.g. ESRI Model Builder
• Climate Change scenarios
• Where a number of possible routes for workflows
  are possible depending on the state of certain
  variables

14
Components

• Geoserver
• 52North WPS
• ActiveBPEL Open Source Engine
• Globus Toolkit (GT4)

or Apache Axis 2
15
SAW-GEO Architecture
16
Geoportal

• User
• Browser-based BPEL Upload
• Hot deployment
• AJAX-based SOAP client
• Workflow Monitoring
• ActiveBPEL Admin Console
• Offers views of deployed processes
• Exports a list of available services to the SOAP
  client as XML
• Red font Developed for SAW-GEO

17
ActiveBPEL Workflow Engine
18
Discussion

• Importing OGC XSD into SOAP services offered high
  interoperability
• However, variations in how tiers interpreted the
  XSD
• varying namespaces
• cyclic imports
• Necessary to throw faults at SOAP tier from
  faults at OGC tiers, because BPEL can handle SOAP
  faults
• Variations in GML between workflow nodes,
  potentially requiring shim services when
  chaining OWS

19
Conclusions

• SOAP wrappers needed for Grid-enabling OGC web
  services
• OGC XSD appropriate for creating SOAP wrappers
• Geospatial data best passed through URL
  references to avoid overwhelming the workflow
  enactor
• BPEL can support Grid-enabled OGC web services

20
Future Work

• Experimental deployment on the Belfast e-Science
  Centre (BeSC) DEBUT system
• Investigate deployment on upcoming NGS GT4
  infrastructure
• User Evaluation (November-December 2007)
• Volunteers needed Please register on the
  website
• http//www.edina.ac.uk/projects/seesaw
• Integration with SEE-GEO