Research On Cooperative Spatial Information Workflow And Its Implementation Technology In Grid GIS

With the rapid development of information processing techniques and the continuous upgrading of computers, the way for spatial information acquisition, production, processing and storage had made great progress. This makes the amount of spatial information in every spatial information application domain increase exponentially. The classic centralized processing method by one single computer has been replaced by the distributed processing mode based on the spatial information infrastructure built on computer networks. This processing model improves the capability of parallel processing and distributed sharing of spatial information. The establishment of current spatial information infrastructure for industry provides high speed network connection for multiple spatial information service nodes. This makes it possible to transform the centralized business processing model into the distributed parallel processing model. Actually lots of online service systems were built on classic Web GIS platform in spatial information infrastructure environment:it doesn't effectively tackle the performance issues and reduce the complexity brought by the classic centralized processing model. The problems are mainly reflected in the following:multiple service nodes can't achieve cooperative computing automatically under the architecture of Web GIS platform. The reality is that every service node operates on their local information resource respectively, and then transfers their processing results to one aggregation node one by one. If the processing strategy changed, updating must be performed on every work node. This aggravates the difficulties of system maintenance and increase the cost. It is impossible to achieve the real distributed spatial information processing.Geospatial Grid Service Workflow System is the emphases of research and development direction of the combination of grid workflow technologies and spatial information processing technologies. It transformed the processing capabilities of a single computer into the Grid processing capabilities by using Grid spatial information processing middleware to shield the complexity of the bottom multi-source heterogeneous spatial information computing resource, then to use Grid services to encapsulate them into the upper abstract uniform resources. After that, it can achieve the real distributed processing of spatial information by service combination and process orchestration. It can fulfill the performance advantages provided by the computer networks. Cooperative Grid spatial information workflow can freely combine GIS functional services and non-GIS-functional services that exist in Spatial Information Grid. By using the business process modeling techniques and the scheduling mechanisms of distributed workflow engine, Cooperation work among working node can be effectively implemented. The efficiency is greatly raised without manual intervention. It also has the advantages of security and fault tolerance of spatial data, and can be integrated with legacy business systems. Through the spatial information workflow management middleware, clients can use many kind of spatial information resource. It improves operating efficiency of spatial information service and provides the reliable technology guarantee for the integration and sharing of spatial information resource distributed among different nodes.This paper, which is supported by the Major Project of National High-tech R&D Program (863Program)"Grid GIS and Its Major Applications", mainly discussed the application model and implementation techniques of spatial information service cooperative processing in Grid computing environment. It focuses on solving the critical problems on spatial information organization and management, Grid-enabled spatial information service modeling, business process construction and computing resource cooperation dynamically. It has practical significance for changing the application patterns of Spatial Information Grid and furthering the combination of Grid Computing and spatial information processing. Through studying on the hierarchical publication and management mechanism for spatial information resource in Spatial Information Grid, a uniform organization and dynamical mapping model is presented. In this model, all multi-source heterogeneous spatial information computing resource, including spatial data, GIS service, are seamless integrated into the resource framework through the cross-platform GIS kernel component and provided to the workflow system modeling layer as Grid Uniform Resource Locator (URL) form. Based on that, workflow framework does process orchestration without considering the complex detail about bottom resources. The service-oriented modeling method is also studied; a Grid service composition model is presented from the perspective of its bottom function-realization. This model has characteristics of easy-disassembly, dynamical-transfer-enabled, rapid assembly. Cooperative spatial information workflow formalization model is provided on the basis of service composition model and a service seamless migration algorithm is also proposed. A series of key implementation technologies of distributed spatial information workflow system were studied:through modifying the classic Grid monitoring and discovering system, a efficient process resource monitoring and discovering system was realized and it can raise workflow system efficiency. Grid service seamless migration function in the process of executing a workflow instance is implemented on the basis of a service composition function supported by the spatial function warehouse. A service integration and invocation tool, as a loosely extension of workflow engine service is realized to support the interaction between workflow system and other legacy service system. Analyzed the application status and business nature of Spatial Information Grid such as China Geology Survey Information Grid (CGSIG) and developed the CGSIG workflow prototype system. Based on that, some workflow application module is rapid implemented and deployed on CGSIG. All this provided a feasible scheme for distributed spatial information retrieval, management, processing, integration and sharing across different industries and departments using Grid spatial information workflow technologies. The details of our study are listed as follow:(1) Analyzed the research status and development trend of distributed computing techniques such as Grid computing and Grid GIS, pointed out the distributed spatial information processing techniques supported by the Grid computing and Cloud computing will be the inevitable development trend of spatial information integration and sharing application. Based on the research status of Grid workflow techniques in spatial information processing field, pointed out that there is a considerable gap between Grid workflow and the current application requirements, and the implementation of cooperative spatial information workflow may better achieve fusion of Grid computing techniques and spatial information field. A series of critical technical problems including process resource organization and scheduling, GIS functional service composition and interoperability must be resolved for the spatial information workflow system.(2) Analyzed the aggregation publish model of spatial information resource on one process node in Grid. Designed the cross-platform GIS kernel framework and realized the integration and uniform publish of multi-source heterogeneous spatial data resource and spatial function resource using spatial data warehouse and spatial function warehouse techniques. Node resource abstract model for spatial information work flow was presented by abstracting the process node resource into four logical resources. Grid service based encapsulation techniques for node logical resources is also discussed. Node resource middleware including node information service, Node resource management service, node status monitoring service and node GIS service was designed and implemented. Based on the relational model of node resource objects, the algorithms of the web interface of node information service, resource management service and node status monitoring are addressed. The node resource model embodies the feature of independent and autonomous, and application-oriented collaborative.(3) Analyzed the global resource model of Grid spatial information workflow. Classified node object, spatial data, GIS service, and spatial virtual organization and Grid users in the global resource directory view by relational model and illustrated the uniform representation for these objects. The user access privileges mapping model for workflow resources was focused on, and some optimization have been done for this model using relational algebra. Based on the global resource directory model, the global resource directory management service was designed to support the dynamical updating mechanism. A three level distributed spatial index model which consist of node, spatial virtual organization and global object is presented to realize rapid resource locating and discovering in global resource or in specific virtual organization. The global redundancy replica model for spatial data is established to support the spatial data replica dynamical match during the process execution, and it improves fault-tolerance for workflow system. Spatial information virtual organization model is discussed and user access privileges in spatial information virtual organization and global range are unified.(4) Analyzed service composition model on the basis of service function template. Based on that, activity network graph is adopted to establish the cooperative spatial information workflow. Formalization of the resource scheduling and migration algorithm was presented. The overall implementation framework of cooperative spatial information workflow is presented. Centering on this framework, the process resource dynamical locating&integration and the implementation of function template based service composition are discussed. the key technologies of distributed service workflow engine was analyzed. The assembly and integration function between workflow system and heterogeneous legacy system is implemented. A three-step Grid service migration model which consists of resource locating, function migration and service template migration was presented and the automatic migration function was realized by using spatial function warehouse and spatial data warehouse techniques.(5) Analyzed key technologies of the cooperative spatial information workflow and application system integration. Establish the overall framework of the collaborative spatial information workflow technology-based application system. The underlying functionality support modules of workflow applications and business organizational model are discussed. By analyzing existing Spatial Information Grid architectures, the Geological Survey Information Grid workflow prototype system was built on the basis of seamless integration of Grid GIS environment and Web GIS-based spatial information service system and the main function modules of the system were illustrated. Choose national important mineral resource potential prediction and evaluation application as research object, design and implement the distributed mineral resources reserves volume-estimating-method-based computing application modules using cooperative workflow techniques. Through the analysis of the results with existing online Web GIS based business systems. The rationality and validity of collaborative spatial information workflow were validated.