| With the development of database technology and internet, the amount of data inrelational databases increased rapidly. People are becoming more and more urgent to sharethe data and information in databases distributed in different geographical locations.Integration of grid databases plays an important role in grid environment for sharing andaccessing various kinds of data resources. It has been paid more and more attention in thescientific computing and commercial application field. A series of studies the methodsbased on the OGSA-DAI system structure can effectively solve integration ofheterogeneous information in the traditional data grid applications. But another importantissue in data integration is to solve semantic heterogeneity among data sources. Semanticmethods for data integrating will become the research focus of data integration technologyin the future. Therefore the research on integration methods oriented to semantic amongGrid Databases has important significance.Semantic Web technologies, so that the computer can "understand" from semanticmeaning of the data, and the data and information can be better processed and used basedon "understanding". Ontology expresses precise meaning of concepts by strict definitionsand relationships of concepts, so as to represent common recognition, shared knowledgeand solve the semantic heterogeneity of the heterogeneous data integration. In this paper,our researches focus on ontology-based semantic integration of heterogeneous databases ingrid environment. Through our studies on a pair of related issues, we put forward ascientific and feasible solution. The main contents of this paper include:(1) We carried out analysis and research on domestic and international theoretical basis,key technologies, associated works and literatures about semantic integration based onontologies. As basic researches in this paper, we analyzed the development of the SemanticWeb and ontology, application status and prospects, elucidated the roles and descriptionmethods of ontology and its evolution, proposed the ontology-based data integrationmethods, discussed semantic processing technologies of the database informationintegration, analyzed concepts of semantic and semantic strength distribution, discussed andanalyzed researches and applications issues of semantic integration of data. Pointed out thatsemantic data integration is feasible by combining fuzzy theory and semantic webtechnologies.(2) Reasoning according to the semantic information provided by the ontologies canmake the user’s queries have certain intelligence. It has more advantages compared with thetraditional methods. Using rich semantic expression capability provided by ontologies, weproposed the overall idea of ontology-based data integration mechanism. In order to avoidthe defects of existing XML, a method for integrating XML data source integrated methodwas proposed which uses ontology as the global schema. Combined with flexible andpowerful data description capabilities of XML, at the same time, it provides an effective method of data integration for data grid.(3) We studied the expansion of ordinary relation to the fuzzy relation and appliedfuzzy OWL ontology to the fuzzy semantic query transform for the relational databases.According to the principles and methods of database query operators in relational schemaand implementation of semantic query, the method that the semantic SPARQL queryreplaces SQL queries over traditional relational databases is developed and its steps andprocess are described in this paper. It is proved that the SPARQL query can replace the fivebasic query operations of relational algebra. The expected results can be obtained by usingthis method. It shows that SPARQL query is complete in the relationship and can simulateany semantic relational query made by the five basic relational algebra operatorscombination.(4) Data integration technology is now developing from static and tight integration tothe dynamic and loose coupling integration. Web services have the characteristics ofautonomy, dynamic, loosely coupled such. Studying on using Web services in the datasemantic integration, we proposed a Web services-based semantic integration model andpresented the implementation method and steps of the system. With hierarchical systemarchitecture design, the system consists of three layers, namely, the presentation layer,functional services layer and storage layer. The system has the characteristics of dynamic,loosely coupled, and has stronger practicability and flexibility. Ontology mapping is theprerequisite and foundation of semantic integration; it can eliminate the heterogeneouscharacteristics of the ontology. So we proposed a Web services-based ontology mappingmethod according to the characteristics of Web services. Based on the principle of the Jenainference engine model, we designed an ontology reasoning subsystem.(5) Based on the above research, combined with a specific campus informationtechnology application cases, designed a data integration solution named digital campusbased on database and its fuzzy semantic. Proposed the master plan for the digital campusinformation platform based on services cloud architecture. Designed a framework based onrelation database for semantic integration of digital campus. For applications data queries,the design given fuzzy ontology creation, storage and management strategies as well asfuzzy semantic query processing solutions to improve queries recall and precision. |
