1 GISolve – TeraGrid GIScience Gateway Shaowen Wang Department of Geography and Grid Research & educatiOn group @ ioWa (GROW) The University of Iowa May.

11

GISolve – TeraGrid GISolve – TeraGrid GIScience GatewayGIScience Gateway

Shaowen WangShaowen Wang

Department of GeographyDepartment of Geographyandand

Grid Research & educatiOn group @ ioWa (GROW)Grid Research & educatiOn group @ ioWa (GROW)The University of IowaThe University of Iowa

May 24, 2007May 24, 2007

22

PurposePurpose

Present how cyberinfrastructure-Present how cyberinfrastructure-based Geographic Information based Geographic Information Systems (GIS) functionSystems (GIS) function

Illustrate how GISolve is developed Illustrate how GISolve is developed to help advance GIScience using to help advance GIScience using cyberinfrastructurecyberinfrastructure

Demonstrate science impact of Demonstrate science impact of GISolveGISolve

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

33

BackgroundBackground

Geographic information quantityGeographic information quantity– Ever increasingEver increasing

Application drivenApplication driven– GPS, LBS, RSGPS, LBS, RS

Computationally intensive geographic analysisComputationally intensive geographic analysis– Heuristic searchHeuristic search– SimulationSimulation– OptimizationOptimization– Statistical methodsStatistical methods

CyberinfrastructureCyberinfrastructure– Grid computing and virtual organizationGrid computing and virtual organization

Open Grid Service Architecture (OGSA)Open Grid Service Architecture (OGSA)– Problem solving environmentsProblem solving environments

Grid portalsGrid portals– TeraGrid Science GatewayTeraGrid Science Gateway

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

44

GISolve Architecture

Computational domain modeling

Domain decomposition

Task scheduling

Information broker and resource discovery

Data access module

Problem solving environments implemented using Grid portal technologies Monitoring

servicesProtocols and services for data access on the

Grid, such as the Globus GridFTP

Grid Middleware such as Globus and Condor

Resource management

55

Spatial Computational Spatial Computational Domain TheoryDomain Theory

Wang, S., and ArmstrongWang, S., and Armstrong, M. P. , M. P. 2005. “A Theory of the Spatial 2005. “A Theory of the Spatial Computational Domain.” In: Computational Domain.” In: Proceedings of GeoComputation Proceedings of GeoComputation 20052005 (CDROM), Ann Arbor, MI, (CDROM), Ann Arbor, MI, August 1-3, 2005August 1-3, 2005

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

66

Information Broker Information Broker and Resource and Resource DiscoveryDiscovery

Self-Organized Grouping method for Grid Self-Organized Grouping method for Grid resource discoveryresource discovery– Padmanabhan, A., Wang, S., Ghosh, S., and Briggs, Padmanabhan, A., Wang, S., Ghosh, S., and Briggs,

R. 2005. “A Self-Organized Grouping (SOG) Method R. 2005. “A Self-Organized Grouping (SOG) Method for Efficient Grid Resource Discovery.” In: for Efficient Grid Resource Discovery.” In: Proceedings of the Grid 2005 WorkshopProceedings of the Grid 2005 Workshop, Seattle, , Seattle, WA, November 13-14, 2005, IEEE Press, pp. 312-317 WA, November 13-14, 2005, IEEE Press, pp. 312-317

Modular Information Provider to support Modular Information Provider to support interoperable information brokeringinteroperable information brokering– Wang, S.Wang, S., , Shook, E., Padmanabhan, A., Briggs, R., Shook, E., Padmanabhan, A., Briggs, R.,

Pearlman, L. 2006. “Developing a Modular Pearlman, L. 2006. “Developing a Modular Information Provider to Support Interoperable Grid Information Provider to Support Interoperable Grid Information Services.” In: Information Services.” In: ProceedingsProceedings of of Grid and Grid and Cooperative Computing - GCC 2006: The Fifth Cooperative Computing - GCC 2006: The Fifth International ConferenceInternational Conference, IEEE Computer Society, pp. , IEEE Computer Society, pp. 448-453 448-453

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

77

GISolve GISolve WorkflowWorkflow

88

A Diagrammatic A Diagrammatic Example of Task Example of Task SchedulingScheduling

1, 1 0, 0 2, 4

4, 2 5, 3

11, 13

14, 14 15, 15

3, 5

6, 6 7, 7

8, 8 9, 9 10, 12

12, 10 13, 11

1, 1 0, 0 2, 4

4, 2 5, 3

11, 13

14, 14 15, 15

3, 5

6, 6 7, 7

8, 8 9, 9 10, 12

12, 10 13, 11

1, 1 0, 0 2, 4

4, 2 5, 3

11, 13

14, 14 15, 15

3, 5

6, 6 7, 7

8, 8 9, 9 10, 12

12, 10 13, 11

1, 1 0, 0 2, 4

4, 2 5, 3

11, 13

14, 14 15, 15

3, 5

6, 6 7, 7

8, 8 9, 9 10, 12

12, 10 13, 11

Small Capacity Large Capacity Medium Capacity

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

99

GISolve Portal GISolve Portal http://www.gisolve.orghttp://www.gisolve.org

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

1010

Domain Decomposition Domain Decomposition PortletPortlet

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

1111

Scientific Analyses Scientific Analyses Supported by GISolveSupported by GISolve

Bayesian geostatistical modelingBayesian geostatistical modeling– Yan, J., Cowles, M. KYan, J., Cowles, M. K., Wang, S., and Armstrong, M. P. ., Wang, S., and Armstrong, M. P.

(2007). Parallelizing MCMC for Bayesian spatiotemporal (2007). Parallelizing MCMC for Bayesian spatiotemporal geostatistical models. geostatistical models. Statistics and ComputingStatistics and Computing, in , in presspress

Detection of local spatial clusteringDetection of local spatial clustering– Wang, S., Cowles, M. K., and Armstrong, M. P. Wang, S., Cowles, M. K., and Armstrong, M. P.

(2006) Grid computing of spatial statistics: using the (2006) Grid computing of spatial statistics: using the TeraGrid for TeraGrid for GGi*(d)*(d) analysis.  analysis. Concurrency and Concurrency and Computation: Practice and ExperienceComputation: Practice and Experience, under revision , under revision

Inverse distance weighted interpolationInverse distance weighted interpolation– Wang, S., and Armstrong, M. P. (2003) A quadtree Wang, S., and Armstrong, M. P. (2003) A quadtree

approach to domain decomposition for spatial approach to domain decomposition for spatial interpolation in Grid computing environments. interpolation in Grid computing environments. Parallel Parallel Computing,Computing, 29 (10): 1481-1504 29 (10): 1481-1504

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

1212

Bayesian Bayesian Geostatistical Geostatistical ModelingModeling

Bayesian geostatistical models based on Bayesian geostatistical models based on Markov chain Monte Carlo (MCMC)Markov chain Monte Carlo (MCMC)– Large-scale spatial-temporal data mining

and inference– Using geostatistical models to characterize

the spatial distributions of environmental processes or disease-related outcomes

Challenges– Computationally intensive

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

1313

Science Impact – A Lung Cancer Science Impact – A Lung Cancer Risk Study Based on Risk Study Based on Residential Radon Residential Radon ConcentrationsConcentrations

Iowa Radon Measurements Predicted Radon Concentrations

From Dr. Brian J. SmithBrian J. Smith

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

1414

Science Impact – Social Science Impact – Social Complexity and the Management Complexity and the Management of Ecosystemsof Ecosystems

A Study on Yellowstone’s Northern Elk Herd From Dr. David A. Bennett

1515

Big Picture – GISolve as an Big Picture – GISolve as an Integrated GIScience Integrated GIScience

WorkbenchWorkbench

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

1616

Education and Education and OutreachOutreach

ModulesModules– SpatialGenSpatialGen

A tool to generate simulated datasets with configurable A tool to generate simulated datasets with configurable geographic distributionsgeographic distributions

– KrigingKriging Spatial predictionSpatial prediction

In classrooms In classrooms – The University of Iowa, 2007The University of Iowa, 2007

Foundations of Geographic Information Systems Foundations of Geographic Information Systems (undergraduate)(undergraduate)

Principles of Geographic Information Systems (undergraduate Principles of Geographic Information Systems (undergraduate and graduate)and graduate)

Bayesian Statistics (undergraduate and graduate)Bayesian Statistics (undergraduate and graduate) Computing in Statistics (undergraduate and graduate)Computing in Statistics (undergraduate and graduate)

– The University of Illinois at Urbana-Champaign, 2007The University of Illinois at Urbana-Champaign, 2007 Geographic information Science (undergraduate and graduate)Geographic information Science (undergraduate and graduate)

Upcoming TeraGrid07 student competitionUpcoming TeraGrid07 student competition

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

1717

ConclusionsConclusions

GISolve demonstrates how GISolve demonstrates how cyberinfrastructure can benefit cyberinfrastructure can benefit research on computationally research on computationally intensive geographic information intensive geographic information analysesanalyses

GISolve integrates OGSA-based GISolve integrates OGSA-based Web services to support the Web services to support the computational aspects of GIServicescomputational aspects of GIServices

Background

Design

Science

Implementation

E&O

Conclusions

Purpose

1818

Ongoing WorkOngoing Work

Extension of the types of geographic Extension of the types of geographic information analyses GISolve supportsinformation analyses GISolve supports– Science impactScience impact

Interoperability of GISolve servicesInteroperability of GISolve services Adaptive domain decomposition servicesAdaptive domain decomposition services Visualization servicesVisualization services Evaluation of GISolve performanceEvaluation of GISolve performance Education and outreachEducation and outreach

1919

AcknowledgmentsAcknowledgments

The University of IowaThe University of Iowa– The Office of Vice President for ResearchThe Office of Vice President for Research– The Office of ProvostThe Office of Provost

The NSF TeraGridThe NSF TeraGrid The NSF ITR: iVDGL projectThe NSF ITR: iVDGL project The Open Science GridThe Open Science Grid My colleaguesMy colleagues

Dr. Marc P. Armstrong (UIowa, GIScience) Dr. David A. Bennett (UIowa, GIScience)Dr. Mary Kathryn CowlesMary Kathryn Cowles (UIowa, Statistics) Mr. Thomas Hansen (UIowa, CS)Ms. Wenli He (UIowa, CS) Mr. Peter Likarish (UIowa, CS)Mr. Yan Liu (UIowa, CS & GIScience) Dr. Brian J. Smith (UIowa, Biostatistics)Mr. Eric Shook (UIowa, GIScience) Dr. Edward WalkerEdward Walker (UT-Austin/TACC, CS) Dr. Jun Yan (UIowa, Statistics) Mr. Junfeng Zheng (UIowa, CS)