|Coordinates||Series A, No. 8|
Web Services, Open Standards, and Advances in Interoperability:
A Selected, Annotated Bibliography
Persistent URL for citation: http://purl.oclc.org/coordinates/a8.htm
of Publication: January 15, 2010
Revised: March 9, 2010
Cynthia Dietz (e-mail: firstname.lastname@example.org) is Science/Map Librarian at the University of Stony Brook (SUNY), Stony Brook, New York.
Abstract: This paper is designed to help GIS librarians and information specialists follow developments in the emerging field of geospatial Web services (GWS). When built using open standards, GWS permits users to dynamically access, exchange, deliver, and process geospatial data and products on the World Wide Web, no matter what platform or protocol is used. Standards/specifications pertaining to geospatial ontologies, geospatial Web services and interoperability are discussed in this bibliography. Finally, a selected, annotated list of bibliographic references by experts in the field is presented.
Keywords:Web services, geospatial data, data processing, spatial data infrastructures, metadata, ontology, interoperability, geoinformatics, grid computing
Begin Page 2
In the last few years, the fields of geospatial visualization of and geoprocessing have grown dramatically because of:
Geospatial Web services (GWS) use language(s), vocabulary (ies), message styles, and program coding to publish objects on the Web. Most geospatial Web services adhere to geospatial standards, developed primarily by the International Standards Organization/Technical Committee 211 (ISO/TC211), the Open Geospatial Consortium (OGC), and the Federal Geographic Data Committee (FGDC) (FGDC 2009), which promotes the National Spatial Data Infrastructure and has endorsed several open standards. Current geospatial standards have been extended by standards developed by the World Wide Web Consortium (W3C), the Internet Engineering Task Force (IETF) (http://www.ietf.org) and the Organization for the Advancement of Structured Information Standards(OASIS) (http://www.oasis-open.org/home/index.php).
Begin Page 3
Just a decade ago, information specialists, including Geographic Information System (GIS) and Map librarians, did well to understand the theory behind geoprocessing and technical issues pertaining to GIS software. Today, they will serve their clients/patrons better if they maintain an understanding and/or proficiency in several additional areas, including: semantic issues introduced by geoprocessing, geospatial Web service interfaces, open source tools, open standards, and Web 2.0.
Most of the resources selected for this annotated bibliography have been published within the last three years. They were selected from subscription databases, WorldCAT, and the Internet. These works, many by renowned researchers, are meant to provide an introduction to developments and research topics in the field.
The resources below have been organized into three sections, the first two of which are “Introduction to Geospatial Web Services”, and “Standards for Geospatial Web Services.” The third section, “Geospatial Web Services,” highlights approaches to improve interoperability, approaches to improve geoprocessing, GWS integrations with mass market applications, and GWS integrations with e-infrastructures. Except where noted, resources within each section are arranged alphabetically. Each work is annotated to include details pertaining to scope, special features, and author affiliation. A complete alphabetical list of all materials cited in this article is located at the end of this bibliography.
This section describes four resources that introduce GWS to readers. They are well written and touch on concepts discussed throughout this article.
Di is the Director of the Laboratory for Advanced Information Technology and Standards (LAITS), a branch of the Center for Spatial Information Science and Systems (CSISS) at George Mason University. He discusses a framework, tested in GeoBrain, for building intelligent geospatial knowledge systems in a Web environment. He asserts that the proposed framework should fully automate the geoquery and geo-assembly steps in geospatial knowledge discovery, fully automate the geocomputation step in limited geospatial domains, and facilitate complex geoprocessing and modeling.
Geo-objects and geo-trees have been developed to characterize geoprocessing functions, which are key to a geospatial knowledge system. A geo-object consists of data, a set of attributes, or a set of methods. A geotree is a processing workflow, with the root formed by geo-objects. Several concepts about geo-objects and geo-trees are introduced. A service chain is a workflow process that could be represented by a geo-tree. It requires a geospatial processing model, a geoprocessing algorithm, a geospatial service module, an archived geo-object and a user geo-object.
Key concepts critical to geospatial knowledge systems are described including data transformation and subsetting, domain-specific ontologies, service catalogs, and query interfaces.
Both a common data environment, such as specified in the OGC WCS, WFS, WMS and Catalog Services, and a common service environment are required so that users’ requests for geospatial data and services can be completed. If extended or profiled for the geospatial domain, Web services from the W3C environment, including Web Service Description Language (WSDL), SOAP, XML and HTTP, may be part of the solution to create a geospatial knowledge system. The prototype GeoBrain tests the functionality of the approach. GeoBrain has been described also by (Zhao et al., 2009) and (Han, Di, Zhao, Wei and Li, 2009).
Begin Page 11
This article develops concepts introduced in Yue et al. (2006). The authors, all from the Center for Spatial Information Science and Systems at George Mason University, promote the use of geospatial Web ontologies with Web services so that a user may exchange geospatial data and information, and execute programs over a network dynamically, quickly and efficiently. In a use case, they propose a new service composition operation in the service-oriented architecture (SOA). In particular, they design and implement the automatic composition of geospatial Web service chains using “Data Type” and “Service Type” ontologies to generate a landslide susceptibility index value. As designed, the “composer” application tool built using the OWL-S composite process automatically combines services into a dependent series and executes the chain. The wrapping of computation services serves as a generalized design for use by a myriad of models beyond the landslide susceptibility product.
Individual services to be chained automatically by the registry are identified: landslide susceptibility, slope, slope aspect, ETM NDVI (enhanced thematic mapper normalized difference vegetation index), WICS and WCS. The data to be chained automatically, where appropriate, would be DEM (digital elevation model), and images (e.g., ETM and Near Infrared (NIR)).
High-level domain ontologies pertain to objects, events or the geospatial domain. For automatic knowledge discovery (Begin Page 12) by the catalog service/composer, the authors propose to bridge application ontologies to high-level domain ontologies using three developed ontologies: “Data Type”, “Service Type”, and “Association”. Here the “Data Type” and “Service Type” ontologies pertain to data types (e.g., terrain elevation and topology) and services (e.g., image processing and data analysis and visualization), respectively. The relationship between services and data, whether direct or relaxed, is described by “Association” ontologies.
To chain Web services having heterogeneous interfaces and messages, schema match tools are required. In many cases, those tools involve reasoning rules related to relationships of class hierarchy. The composer tool sets the match options for the “Data Type” and “Service Type”. The mediated RDF framework registers mappings in the grounding information of OWL-S.
See Alameh (2003) and Lutz et al. (2009) for discussions on mediation .
The authors, affiliated with the European Commission, and other geomatics centers in Germany, present an ontology-based method involving an ontology architecture, an ontology language, and reasoning procedures to improve interoperability in spatial data infrastructures (SDI). They introduce two scenarios, from the geology and hydrology domains, to illustrate the benefits of such a method for the discovery, retrieval, interpretation, and integration of information at three levels: metadata, schema and data content. They also show how the method can be encapsulated in services and client applications.
In the geology scenario, the scientist is required to formulate a Styled Layer Descriptor. In the hydrology scenario, the scientist is attempting to implement a Web service chain.
To facilitate the retrieval of relevant sources, ontologies, rather than queries, are used in search and retrieval processes to avoid the need to get user confirmation. Specifically, the approach uses subsumption relationships to build ontologies, allowing no gradual differentiation. The method contrasts with feature-based approaches to ontology used in the OGC Interoperability Experiment.
In the hybrid ontology approach described, a shared vocabulary of a common domain is built using concepts from each application. The selected knowledge representation language is a Description Logic (DL) notation, the basis of OWL. Tests for subsumption reasoning are used for “matchmaking”. Other tools are used as well (e.g., mediation, semantic descriptions involving two parts, and context transformation rules) to build the model to ensure that terms used are explicit and comparable.
To produce consistent results, the methods described above are encapsulated in software components. They support dynamic service chaining. In all cases, an Ontology-based Reasoner, and a Client Workflow Service (which acts as a Web service client), are used. If contextual heterogeneity persists, other components are activated: a semantic translation specification service (for deriving a transformation), a translation service (for executing the transformation), and an interpolation service (for interpreting WFSs data output).
Related projects (BUSTER, SEWASIE, SEEK, GEON and HarmonISA) are discussed and compared to the proposed architecture. In the future, more complex and heterogeneous scenarios will be tested, as well as those where “scale” matters. They will address complications in the semantics of services, the use of templates to describe service chains, and formal geodata description automation.
The authors, from the Center for Spatial Information Science and Systems at George Mason University, propose a taxonomy of geospatial services so that services may be discovered by service category and version. The system captures information about service interfaces, and their inputs and outputs. Its strength lies in two aspects: it facilitates accurate service discovery, and describes how the service may be reused. Use of the taxonomy for the Global Earth Observation System of Systems (GEOSS) is introduced.
The authors discuss service taxonomies used by ISO, NASA’s GCMD, and OGC. None provide the discovery capabilities of the taxonomy proposed. The model proposed is hierarchical with six layers: service category, service type, version, profile, binding, and uniform resource name (URN). Each classification node, identifying one classification concept, may be identified by its position in the classification tree, or by an URN. The use of such a taxonomy facilitates discovery, evaluation for fitness, and dynamic integration.
There appears to be no equivalent taxonomy to facilitate global service discovery for geospatial services based on the interoperability of interfaces. Still, the proposed service taxonomy is limited in that it does not represent service content. The system would benefit from user evaluation and feedback.
When geospatial Web services, Web services and ontologies are used together, as described here, large volumes of data in a scientific work-flow may be analyzed. This article is tough reading for those who have not mastered Web service terminology. The authors use common formats for data interchange and a Web ontology language in combination with a system of Web services and knowledge management technologies, known as GeoBrain, so that users may collaborate to develop executable service chains that produce the products desired. The authors describe geospatial knowledge transformation, a geospatial domain ontology, an ontology-based knowledge base, and a semantically enabled catalog service. They detail steps in geospatial knowledge transformation, and mention related work. For other discussions on GeoBrain, see Di (2005) and Han, Di, Zhao, Wei and Li (2009).
In geospatial knowledge transformation, three phases translate expert knowledge into a data product: geospatial modeling, model instantiation (obeying rules and constraints to generate a workflow) and model execution (to generate data products). To address semantic heterogeneities (e.g., issues of geospatial classification, representation and relationships) and structural heterogeneities (e.g., differences in geospatial data formats, projections, and computing platforms) four ontology models are offered that specify the syntax of geospatial objects, relationships and services: General Ontology, Geospatial Domain-Specific Ontology, Geospatial Data Ontology, and Geospatial Process Ontology.
Dublin Core Metadata provides the core upper-level vocabulary for the general ontology. The Geospatial Domain-Specific Ontology is provided by experts and covers spatiotemporal factors, physical facts, disciplines, and platforms. It provides scientific meanings to data resources.
Geospatial Process Ontology is a model conceptualizing service types: it depicts feature processes, classes them, and documents relationships and constraints. Concepts of methodology, algorithms, and input-output are incorporated. A semantically enabled OGC CSW, and an ebRIM profile, are used to discover and access the geodata.
In geospatial knowledge transformation, the domain expert sets a goal, finds a service type, and registers the model. Web services may be classified as a geospatial process service (e.g., OGC WPS), a geospatial fusion service, or a geospatial data service (e.g. WCS, WFS, and WMS). In GeoBrain, the Business Process Execution Language for Web (Begin Page 14) Services (BPEL4WS) is used to represent the service chain, because model reusability is critical. In the service chain, service types are mapped, geospatial data services are joined, and geospatial fusion services make the data discoverable. Composite service bindings execute the service chains.
The authors, from the Center for Spatial Information Science and Systems at George Mason University, mention several workflow systems designed to execute end-to-end processes, such as Taverna, Kepler, and SciFlo. They conclude that research is needed to include spatial reasoning in Web semantics. They hope to research rules appropriate for spatial inference for improved data and service discovery.
Alameh, associated with Global Science & Technology, Inc. describes the chaining of GIS Web services. Dynamic access to customized geographic information is becoming possible, given the availability of specialized interoperable geospatial Web services, the availability of spatial data, the applicability of the data to location-based services, and models that give users just the data they need and not more.
Geospatial Web services, which can be invoked, located or published by users, may be described in three groups: data services, processing services and registry or catalog services. They are accessed via standard protocols, including HTTP and SOAP.
Complementary services may be chained to create custom applications. The author graphically illustrates service chaining, as well as the architecture of GIS Web services. Service chaining falls into three groups: client-coordinated, static or mediated. Details of each chaining method are described well. Mediated services tend to be complex, and oriented to a specific domain. Working dynamically, they may yield inconsistent results.
DAML-S, now known as OWL-S, is described as one of several XML technologies that can support geospatial Web services, and it is perhaps the best for service chaining.
The author, from the University of Bonn, introduces a WPS, now an approved by OGC service, that uses grid data to generate just-in-time access and information from distributed geodata inventories. Two case studies are presented. The role of geospatial Web services (GWS) within the Spatial Data Infrastructures (SDI) is described. The underlying data can be updated without affecting user interaction. GWS provide heterogeneous and distributed access to data of the client’s choice.
SDIs are typically described as having three tiers: Data tier (the backend), the Business Logic tier (an integration tier or middleware) and the Presentation tier (front-end). Within the Business Logic tier, tasks for geoprocessing occur via algorithms and interfaces designed for interoperability. Three of those common interfaces are described: getCapabilities, describeProcess, and execute. Both the get capabilities and describe process interfaces help generate the information needed for complex spatial processes via service chains. Each service pertains to one process, be it a spatial intersection or spatial buffer, which is executed in a certain order in the chain. Service chains, being independent of data and context, may be used on any platform, or by any implementation language.
Begin Page 15
In the first case study, a simple spatial intersection service is developed which uses two GML datasets. One service module—handling processes such as request validation, error handling, and output preparation—has the potential to be reusable. The second service module for spatial intersection was built with an algorithm. Since it required some customization, it is not as reusable.
In the second case study, geoprocessing handled data from 20 different sources that contributed to the generation of a groundwater vulnerability index. The data was represented at a variety of scales: microscale, mesoscale and macroscale. Three steps, taken to estimate overall protective effectiveness, generated “factors” (via OGC compliant services), transformed all data to grid data, computed groundwater vulnerability via an equation and Map Algebra, and provided a SOAP Web service via a WSDL interface. Each step in the process is well explained by the author.
The consumer decides whether the result from the interface should be presented inside a traditional GIS or in a mobile environment. In a WPS, the user defines the area of interest (e.g. via a bounding box), the coverages or features to be used, and the “topics” to be generated by the system. Although the process speeds processing, via just-in-time generation, complex modeling is not done in an automated way. Manual adjustments are required. Still, when a provider updates data, the WPS result is updated the next time the service is accessed.
The authors conclude that the development of WPS interfaces is time consuming, yet valuable, since interoperability results. The model presented uses GWS without excluding W3C-compliant Web services, such as SOAP.
The authors, from lat/lon GmbH, and the universities of Aachen and Bonn, use the OGC’s draft WPS specification in two case studies. From the use cases they derived the technological building blocks of the Web Service Orchestration model, designed to process geodata in an OGC compliant way. In service chaining via the orchestration model, the services are loosely coupled: neither the successor nor the ancestor is exactly known.
WPS permit users to perform geoprocessing tasks, such as the spatial intersection of features, the conversion of vector data to raster data, and buffering using geoprocessing algorithms.
Case study one reflects work described in Kiehle (2006) pertaining to groundwater vulnerability. Although the workflow is not reusable, the services are. The second case study, pertaining to land parcel information, automates the workflow. The authors describe the steps to provide a service-driven automated property information system. In this case, XML-Remote Procedure Calls were used rather than SOAP messaging.
Some problems were encountered in both case studies. In the future, the authors will research flexible chaining of process units. The technical preconditions for multiple WPS service instances to be handled exist, but a model is lacking to semantically describe spatial operations.
The authors have proposed a Web service orchestration framework, calling on several services and processes (Process Repository, Service & Data Registry Services, Rules Repository, Rules Engine, Geodata Access Services, Geodata Manipulation Services, and Geodata Portrayal Services) and an orchestration engine as a central service chaining unit. The workflows and rules are programmed in a logical way.
Begin Page 16
The authors, from universities and geomatics institutes in Germany, integrate geoprocessing with geospatial mass-market applications. Additionally, they demonstrate the capabilities of the approach in a fire threat use case.
Features of the OGC KML standard are described, and its ability to dynamically integrate remote resources. The OGC WPS is also described, and three of its operations: GetCapabilities, DescribeProcess and Execute. Note a similar discussion of WPS in Granell (2008). WPS processes and stores data retrieved from an URL, and may deliver it as raw data, which makes integration with geospatial mass-market applications possible.
In the proposed approach, uDig, a WPS client, exports a KML file to mass-market applications (e.g. Google Earth) using one of two options: static or dynamic. Subsequently a process is referenced or executed.
In a use case scenario, an expert using uDig configures a buffering and intersecting process pertaining to a fire threat scenario, and exports the process in a KML file. A citizen loads the KML into a portal to visualize the results.
The approach presented complies with KML and WPS standards. Interfaces and encodings do not require customization. Research on the integration of more complex process chains is anticipated.
The authors, from the Universitat Jaume I in Castellon, Spain, seek to provide a distributed, scalable and easier approach to workflow pertaining to hydrological models and datasets. The solution presented integrates geospatial Web services with mass market mashup technology for improved visualization. Basic concepts are introduced, the system architecture is described, and the application scenario demonstrated.
A key to providing more complex services is the OGC WPS. WPS is an interface that describes functionalities, and may wrap off-line services as Web services. Three methods describe service functions: getCapabilities, describeProcess, and execute. The ability of WPS to wrap geospatial services with general purpose ones increases interoperability significantly.
Mapping mashups are applications that use a variety of services, and are used frequently for their visualization capabilities. Updates are shown in real time. The mashup relies on a client-side map service(s) and appropriate data access.
The general architecture used is based on the European Spatial Data Infrastructure INSPIRE. Besides the Data layer, three loosely-coupled layers (i.e. Presentation, Horizontal and Service) describe the architecture. The Presentation layer provides access to data and services, the Horizontal layer guides users through interfaces involved in configuring hydrological models and the Service layer groups service instances. The services grouped pertain to processing, downloading, and viewing. Details of operations within each layer are given.
Begin Page 17
The implementation of the hydrological modeling provides access to the specific processes and data needed for each model, saving much time and effort. Details of the implementation steps are given. Processes wrapped by WPS may include raster analysis, spatial intersections and coordinate transformations.
The geoprocessing services developed in the hydrological modeling are reusable and registered in OGC catalogs. Because WPS works with algorithms and not pre-determined datasets, they can be chained to such services as WMS and WCS, providing functionality usually only achieved via desktop packages. GML processes elevation zones, while the mashup integration logic component transforms them into KML. Results are displayed using Google Maps API.
Future research will involve a model engine to orchestrate WPS services for complex scientific models.
The editors, affiliated with the University of Otago, New Zealand and Wilfrid Laueir University, in Waterloo, Ontario, have produced a highly recommended book describing the modeling for and use of Free and Open Source for Geospatial (FOSS4G) software. Many GWS or OGC models are used in combination with FOSS4G software, including: MapServer, MapGuide, GeoTools, GRASS GIS, GeoVISTA Studio, MapChat and TerraLib. The variety of software integrations described indicates how powerful GWS implementations may become in the future.
The authors, from the Center for Spatial Information Science and Systems at George Mason University, describe the design and implementation of GeOnAS, a Web service-oriented online geospatial analysis system that makes NASA’s Earth Observing System (EOS) and other data available to geoscientists for access and modeling. Additionally, the authors present developments in the field, the system architecture, and details of each module. For other discussions on GeoBrain see Di (2005) and Zhao et al. (2009).
In implementing GeOnAS, Web services and AJAX have allowed the developers to maximize analysis, visualization, and modeling capabilities. The general architecture has four layers: the browser client, the interface, services, and database server. Interfaces include modules, such as User Portal, Data Management, Data Visualization, Data Analysis, Catalog, and Workflow. The Services layer includes WCS, WFS, Geographic Markup Language (GML), WMS, Web Map Context (WMC), CSW, and WPS. The database server layer includes the GMU-LAITS, and NASA-ECHO databases.
GeOnAS includes modules for management, manipulation, display, analysis and invocation. The functions of each module are described. The output could be saved, or a KML file created for integration into Google Earth or Google Maps.
The system has displayed superior performance capabilities for publishing, accessing, processing, retrieving, knowledge building, and sharing. Future improvements designed include the support for the Opera and Safari browsers, user-defined geoprocessing, and complex analysis.
Begin Page 18
Several national and international initiatives use grid computing with semantics or ontology tools to advance interoperability including: the Geosciences Network (GEON) (Baru et al. 2009), National Science Foundation’s (NSF) Ocean Observatories Initiative (Arrott et al 2007, and Farcas et al. 2008), and the Virtual Solar-Terrestrial Observatory (Fox et al. 2009). Those that use geospatial Web services are not fully OGC and ISO compliant. Attempts are being made to grid-enable GWS to improve performance and security frameworks. Several exceptional efforts to integrate GWS with Grid computing are presented in a special edition of GIS.Science (http://portal.opengeospatial.org/files/?artifact_id=35975). Some of those and others are described and annotated here. These resources use grid computing technology, or e-infrastructures, in combination with GWS to share geospatial data, computing power, algorithms and/or other methods to handle complex multidisciplinary problems.
Hobona, Fairbairn, Hiden and James, all from the University of Nottingham, Newcastle University and the North East Regional e-Science Centre, U.K., very clearly detail an innovative proposal that uses GWS and tools supported by OGC or Open Grid Services Architecture (OGSA) standards. Their proposal integrates GWS with Grid services to enhance geoscientific workflows and uses workflow enactors to support the orchestration of geoscientific GWS. Workflows are essential to geoscientists to assist in data management, processing, and analysis.
Two enactors are tested with the SAW-GEO (Semantically-Aware Workflow Engines for Geospatial Web Service Orchestration) project: the Simple Conceptual Unified Flow Language (SCUFL) and the Business Process Execution Language (BPEL). See highlights of SAW-GEO in a presentation by Reed (2008).
Workflows are created using several geospatial Web services, including WCS, WFS, WMS and WPS. WMS provide the visualization of WFS and WCS, while WPS provide processing, computational and analytical functions, often performed using algorithms.
OGC and OGSA provide different mechanisms for publishing, finding and binding services: OGC may use GML while OGSA uses SOAP. The authors propose a work-around which involves storing a GML document in a Web-accessible folder, while using a SOAP message containing an URL to reference the document. A WPS transmits feature collections by delivering URL references to GML documents.
The authors also propose a SOAP-based service, which they call a proxy service, to wrap GWS services using SOAP-based interfaces. The interfaces import OGC XML schemas into a Web Services Resource Framework (WSRF) WSDL. Such a proposal when used with other proxy services, servlets and parameters enable the dynamic referencing of target GWS for geopocessing at runtime. The servlet provides access to a dataset during workflow enactment. Possible enactors include ActiveBPEL (a BPEL enactor) and Taverna, which uses SCUFL.
To implement the workflow, Globus Toolkit with a WSRF interface hosted the OGSA services. The open source 52 North WPS and Geoserver were used to contain OGC services. A workflow involving parallel “Thiessen” and “Union” processing and independent subprocessing was set up. Both ActiveBPEL and Taverna separately implemented the workflow. Both implementations were successful, and each had advantages. Both implementations handled large geospatial datasets, OGC data types and parallel sub-processing.
Begin Page 19
Lee and Percivall describe e-infrastructures as platforms which support applications that may use multiple data sources, and which process and consume the data—possibly in multiple locations. Some have strong security models. (Begin Page 20) The OGC and Open Grid Forum (OGF) strive to develop geospatial applications with e-infrastructures, as in the OGC Web Services-Phase 6 Demonstration (OWS-6) (OGC 2009) airport disaster scenario.
Geospatial data needed in e-infrastructures to address complex problems, as presented by environmental monitoring, and energy and disaster management, will come from academic, industry, government, and virtual organization archives and sources in the field. Several projects have incorporated such data, including the German National D-Grid and CYCLOPS, while others are helping to facilitate such efforts, including INSPIRE, EGEE, the European and American Geophysical Unions, the National Science Foundation, and the US Federal Geographic Data Committee.
The authors suggest that both cloud computing by various governments (e.g., Japan’s Kasumigaseki Cloud) and disasters such as Katrina will drive the development and interoperability of geospatial applications and infrastructures. For models to successfully predict the impacts of such disasters as Katrina, they suggest that several fields need to advance, including: atmospheric and oceanic science, computational science, operational infrastructures, and user-friendly geospatial information systems.
Padberg and Kiehle, from the University of Bonn, and lat/lon GmBH respectively, compare the OGC and Grid computing paradigms, and describe how Grid computing can be used in an OGC context. Since many geospatial collections have been centralized, remote geoprocessing, as in an OGC WPS, is now possible. Challenges to implementing a spatial data infrastructure (SDI) with geoprocessing capability exist. Many SDIs have huge amounts of data that the the data originators/owners would like to securely process in an optimum way. Although Grid computing offers high-performance, distributed, large-scale data sharing, none provides a fully compliant SDI infrastructure.
Several characteristics of Grid computing are incompatible with conventional SDIs, including service description documents, service interfaces, "stateful services" (see definition of state in glossary), and security mechanisms. The differences are described. Several prototypes, or use cases, of a Grid-enabled SDI were tested that use the Java framework "deegree"( http://www/deegree.org) to build the SDI, and Globus Toolkit 4 to provide the Grid middleware. In all, the WCS, WFS and WPS specifications were modified. Customizations included creating a custom datastore (a database integrating data from multiple sources), creating a Grid service, inserting WPS logic into the Grid service, the use of grid-specific security settings, and enhancements to communicate with a MyProxy repository.
The modifications described were implemented in the German SDI project, known as GDI-Grid-Project. Beyond an improved capability to store and compute, the project provided Grid users the ability to integrate geospatial service calls into workflows.
Future work will involve the creation and validation of automated Grid workflows, and further integration, generalization and enrichment of data in datastores. Possible use case scenarios are described involving noise propagation, flood simulation, and emergency routing.
Some issues pertaining to grid-enabled SDIs are presented. User authentication tends to delay request cycles. Parallel processing might be difficult if no parallel algorithms are available. Grid computing may delay processing for a few hours, impacting the integration of real-time sensor data.
Possible enhancements could permit a Grid infrastructure to split storage or computing processes, and to execute the subprocesses simultaneously, which would speed processing. Virtual organizations could make use of the system simultaneously from remote locations.
Begin Page 21
Woolf and Nativi are from the Science & Technology Facilities Council Rutherford Appleton Laboratory, U.K., and the Italian National Research Council, Institute of Methodologies for Environmental Analysis, respectively. They describe efforts at developing informatics systems and grid or e-infrastructures to meet the needs of earth science endeavors, which deal with global datasets in a multidisciplinary approach. They argue that e-infrastructures involving modular systems are needed to create persistent services for complex analyses. Service oriented architectures (SOA) now provide the modularity needed to transform data-centric services from “vertical stacks” to service-oriented ones that work through robust registries. Using examples, the authors discuss the challenges for Spatial Information Infrastructures (SII), for advanced earth science grid infrastructures, and for informatics. They cite the need for models and tools based on international standards that incorporate global data from multiple disciplines and describe physical processes having temporal and spatial dimensions. Such models and tools would help those seeking to respond to environmental problems such as climate change and biodiversity.
Two international initiatives and grid infrastructures for Earth science, GEOSS and GMES (Global Monitoring for Environment and Security), are mentioned as needing grid technologies for the sharing of resources in virtual organizations. Grid infrastructures provide improved distributed processing capabilities and are reliable, scalable and secure. The grid application, CYCLOPS, and aspects of NERC DataGrid, a federated data infrastructure, are described. Their efforts to integrate ISO and OGC standards with the grid infrastructures are not complete, and would benefit from best practices. Transitions from file-based to content-based information management are challenging. The GeoSciML and CSML efforts (data models and GML schemas in geosciences (Allison et al., 2008) and climate sciences) are noted for making advances in multidisciplinary modeling.
Coverages and temporal information used heavily in several earth sciences, including oceanography and meteorology, require advanced semantic processing, ontologies, feature catalogs, and the use of observation and measurement schemas and sampling strategies, as offered by Cox (2007a and 2007b). Data simulation models need to be tested using real observations. The authors describe two use cases involving an on-demand flood-risk assessment, and the measurement of a mesoscale eddy.
Several e-infrastructure efforts are noted that require the interoperability provided by geospatial service standardization. Such efforts are not fully interoperable. Substantial progress has been made in the GEOSS Architecture Implementation Pilot (Phases 1 and 2) (http://www.ogcnetwork.net/AIpilot), and in the GEOSS Interoperability Process Pilot Project (IP3), as described above. The authors’ challenge to the earth science community to continue to enhance interoperability efforts is strong, and the call for integrating ISO and OGC standards, with enhancements, clear.
Woolf and Shaon offer an approach to use Grid processing to add value to a WPS so that the geoprocessing and analysis for models and large datasets may be scheduled, enhanced, and secure. OGC geospatial catalogs enhance Grid data movement tools. Grid computing provides a framework needed by GWS to construct complex workflows using several computing nodes if needed. Job Submission Description Language (JSDL), an OGF specification, describes the data and computational resources needed for an implementation. When combined with the WPS process description and data, a valid JSDL can be formed to manage a large computational process. The authors outline the steps to create a Grid-enabled WPS service, which involved the creation of a WPS Grid profile and a WPS SOAP /Proxy layer. They note an implementation in scene 4 of the geoprocessing demonstration in OWS-6 (OGC 2009). Future enhancements are needed, (Begin Page 22) including improved integration and development of security practices, the use of the WSRF, and the use of middleware conforming to the HPC Basic Profile.
Alameh, Nadine. 2003. Chaining geographic information Web services. IEEE Internet Computing 7(5): 22-29 (accessed on August 25, 2009 from IEEE Xplore database ).
Allison, M. Lee, Linda C. Gundersen, Stephen M. Richard, and Tamar L. Dickinson. 2008. Implementation plan for the Geoscience Information Network (GIN). U.S.G.S. Scientific Investigations Report 2008-5172: Geoinformatics 2008-Data to Knowledge, http://pubs.usgs.gov/sir/2008/5172/ (accessed on December 13, 2009).
Arrott, Matthew, Alan Chave, Ingolf Krueger, John Orcutt, Alex Talalayevsky, and Frank Vernon. 2007. The Approach to cyberinfrastructure for the Ocean Observatories Initiative. Paper presented at Oceans 2007, https://sosa.ucsd.edu/ResearchCentral/download.jsp?id=175 (accessed on December 13, 2009).
Bai, Yuqi, Liping Di and Yaxing Wei. 2009. A taxonomy of geospatial services for global service discovery and interoperability. Computers & Geosciences 35: 783-790 (accessed on August 25, 2009 from Academic Search Complete database).
Baru, Chaitan, Sandeep Chandra, Kai Lin, Ashraf Memon, and Choonhan Youn. 2009. The GEON service-oriented architecture for earth science applications. International Journal of Digital Earth 2. Suppl. no.1: 62-78.
Cox, Simon, ed. 2007a. Observations and Measurements. Part I. Observation Schema. http://www.opengeospatial.org/standards/om (accessed on December 13, 2009).
Cox, Simon, ed. .2007b. Observations and Measurements. Part II. Sampling Features. http://www.opengeospatial.org/standards/om (accessed on December 13, 2009).
Di, Liping. 2005. A framework for developing Web-service-based intelligent geospatial knowledge systems. Paper presented at the annual international meeting of the Association of Chinese Professionals in Geographic Information Science (CPGIS), http://www.iseis.cuhk.edu.hk/downloads/full_paper/2005-24-28.pdf (accessed on July 26, 2009).
Farcas, C., P. Fox, M. Arrott, E. Farcas, I. Klacansky, I. Krueger, M. Meisinger, and J. Orcutt. 2008. OOI CyberInfrastructure: Next generation oceanographic research. Poster presented at the annual meeting of the American Geophysical Union, https://sosa.ucsd.edu/ResearchCentral/download.jsp?id=168 (accessed on December 13, 2009).
Federal Geographic Data Committee (FGDC). 2009. The Federal Geographic Data Committee. http://www.fgdc.gov/ (accessed on December 13, 2009).
Foerster, Theodor et al. 2009. Integrating OGC s into geospatial mass-market applications. Paper presented at the GEOWS ’09 International Conference (accessed on December 20, 2009 from the IEEE Xplore database).
Fox, Peter, Deborah L. McGuinness, Luca Cinquini, Patrick West, Jose Garcia, James L. Benedict, and Don Middleton. 2009. Ontology-supported scientific data frameworks: The virtual solar-terrestrial observatory experience.” Computers & Geosciences 35: 724-738, (accessed on November 12, 2009 from Academic Source Complete). Also available online (Begin Page 23) at http://www.hao.ucar.edu/projects/vsto/presentations/C_AND_G2007_final_figs.pdf
Granell, C., L. Diaz and M. Gould. 2008. Geospatial Web service integration and mashups for water resource applications. Paper presented at ISPRS Congress Beijing 2008, Beijing, China, http://www.isprs.org/congresses/beijing2008/proceedings/4_pdf/117.pdf (accessed on July 25, 2009).
Hall, G. Brent and Michael G. Leahy. 2008. Open source approaches in spatial data handling: Advances in geographic information science. Berlin: Springer.
Han, Weiguo et al. 2008. Design and implementation of GeoBrain Online Analysis System (GeOnAS). Paper presented at the 8th International Symposium, W2GIS, Shanghai, China, http://geobrain.laits.gmu.edu/papers/W2GIS2008_GeOnAS.pdf (accessed on August 1, 2009).
Hobona, Gobe, David Fairbairn, Hugo Hiden, and Philip James. Orchestration of grid-enabled geospatial Web services in geoscientific workflows. Forthcoming in IEEE Transactions on Automation Science and Engineering. Available as a preprint from the IEEE at http://ieeexplore.ieee.org/xpl/tocpreprint.jsp?punumber=8856&isnumber=4358066&isYear, doi:10.1109/TASE.2008.2010626.
ISO/TC211. Standards Guide: ISO/TC211 Geographic Information/Geomatics, http://www.isotc211.org/Outreach/ISO_TC%20_211_Standards_Guide.pdf (accessed on January 6, 2010).
Khalsa, Siri Jodha Singh, Stefano Nativi, and Gary N. Geller. 2009. The GEOSS interoperability process pilot project (IP3). IEEE Transactions on Geoscience and Remote Sensing, 47 (1): 80-91 (accessed on December 20, 2009 from the IEEE Xplore database).
Kiehle, Chistian. 2006. Business logic for geoprocessing of distributed geodata. Computers & Geosciences 32: 1746-1757 (accessed on August 25, 2009 from Academic Search Complete database).
Kiehle, Christian, Christian Heier and Klaus Greve. 2007. Requirements for next generation spatial data infrastructures-standardized Web based geoprocessing and Web service orchestration. Transactions in GIS 11:6, 819-834 (accessed on August 25, 2009 from Academic Search Complete database ).
Kralidis, Athanasios Tom. 2007. Geospatial Web services: The evolution of geospatial data infrastructure. In The geospatial Web: How geobrowsers, social software and the Web 2.0 are shaping the network society, edited by Arno Scharl and Klaus Tochtermann, 223-228. London: Springer-Verlag.
Kresse, Wolfgang and Kian Fadaie. 2004. ISO standards for geographic information. Berlin: Springer-Verlag.
Lake, Ron, David S. Burggraf, Milan Trninic, and Laurie Rae. 2004. Geography mark-up language (GML). Chichester: John Wiley.
Lee, Craig. A. and George Percivall. 2009. The evolution of geospatial e-infrastructures, GIS.Science 3: 68-70, http://portal.opengeospatial.org/files/?artifact_id=35975 (accessed on October 16, 2009).
Lieberman, Joshua. Geospatial Semantic Web Interoperability Experiment. http://www.opengeospatial.org/projects/initiatives/gswie (accessed on August 12, 2009).
Lutz, M., J. Sprado, E. Klien, C. Schubert and I. Christ. 2009. Overcoming semantic heterogeneity in spatial data infrastructures. Computers & Geosciences. 35: 739-752 (accessed on August 25, 2009 from Academic Search Complete database ).
Begin Page 24
Moellering, Harold, Henri J. G. L. Aalders, and Aaron Crane. 2005. World spatial metadata standards: Scientific and technical descriptions, and full descriptions with crosstable. San Diego: Elsevier.
Morris, Steven P. 2006. Geospatial Web services and geoarchiving: new opportunities and challenges in geographic information services. Library Trends, 55:2 (Fall): 285-303. (accessed August 1, 2009 from Wilson Web database).
National Science Foundation. 2007. Cyberinfrastructure vision for 21st century discovery, Chapter 4: Virtual organizations for distributed communities (2006-2010), http://www.nsf.gov/pubs/2007/nsf0728/nsf0728_5.pdf (accessed on December 13, 2009).
Nebert, Doug, Carl Reed, and Roland M. Wagner. 2007. Proposal for a spatial data Infrastructure standards suite: SDI 1.0. In Research and theory in advancing spatial data infrastructure concepts, edited by Harlan Onsrud, 147-159. Redlands, CA: ESRI Press.
Open Geospatial Consortium (OGC). Catalogue Service (CAT). http://www.opengeospatial.org/standards/cat (accessed on August 18, 2009).
OGC. OpenGIS Standards and Specifications. http://www.opengeospatial.org/standards (accessed on August 16, 2009).
OGC. OGC Reference Model (ORM). http://www.opengeospatial.org/standards/orm (accessed on December 12, 2009).
OGC. OGC Web Services- Phase VI Demonstration. http://www.opengeospatial.org/pub/www/ows6/index.html (accessed on December 13, 2009).
Padberg, Andrew and Christian Kiehle. 2009. Towards a grid-enabled SDI: Matching the paradigms of OGC Web services and grid computing. International Journal of Spatial Data Infrastructures Research, Special Issue GSDI-11. http://ijsdir.jrc.ec.europa.eu/index.php/ijsdir/article/view/133 (accessed on December 21, 2009).
Reed, Carl. 2008. GRID Computing and the geospatial Web: The role of standards. Powerpoint presented at the Mardi Gras Conference in Baton Rouge, LA, http://www.ogcnetwork.net/ogcpresentations (accessed on December 13, 2009).
Wiederhold, Gio. 1999. Mediation to deal with heterogeneous data sources. Lecture Notes in Computer Science (LNCS) 1580: 1-16 (accessed on December 21, 2009 from SpringerLink database).
Woolf, Andrew and Stefano Nativi. 2008. How earth science can contribute to and benefit from the spatial information infrastructure. In Creating spatial information infrastructures edited by Peter J. M. Oosterom and Sisi Zlatanova, 67-87. Boca Raton: CRC Press.
Woolf, Andrew and Arif Shaon. 2009. An approach to encapsulation of grid processing within an OGC Web processing service. GIS.Science 3: 82-88, http://portal.opengeospatial.org/files/?artifact_id=35975 (accessed on November 2, 2009).
World Wide Web Consortium (W3C). OWL-S: Semantic markup for Web services. http://www.w3c.org/Submission/2004/SUBM-OWL-S-20041122/Overview.html (accessed on August 16, 2009).
Yue, Peng et al. 2006. Semantic augmentations for geospatial catalogue service. Paper presented at the 2006 IEEE International Geoscience and Remote Sensing Symposium. Denver, http://www.laits.gmu.edu/geo/nga/doc/semanticCatalog.pdf (accessed online on August 26, 2009).
Yue, Peng et al. 2007. Semantics-based automatic composition of geospatial web service chains. Computers & (Begin Page 25)Geosciences 33:649-665, http://laits.gmu.edu/~gyu/pdfs/cg2007.pdf (accessed on August 21, 2009)
Zhao, Peishing, Genong Yu and Liping Di. 2007. Geospatial Web services. In Emerging Spatial Information Systems and Applications, edited by Brian N. Hilton, 1-35. Hershey, PA: Idea Group. Also available online at http://www.laits.gmu.edu/geo/nga/doc/geoWebService.pdf
Zhao, Peisheng et al. 2009. Semantic Web-based geospatial knowledge transformation. Computers & Geosciences 35: 798-808 (accessed on August 25, 2009 from Academic Search Complete database ).
Begin Page 26
Begin Page 27
Begin Page 28
Begin Page 29
Begin Page 30
Begin Page 31