Abstract. In allusion to scientific data isomerism and diversity as well as the lack of semantics in petroleum domain, this paper puts forward ontology-based semantic extraction method for heterogeneous data. Firstly, semantic extraction framework for scientific data in petroleum domain is constructed. On this basis, corresponding semantic transformation and abstract rules are worked out in allusion to structured and semi-structured data to form local domain ontology. Then, overall ontology in petroleum domain of data is constructed through dependence mapping among heterogeneous data. Effective organization and uniform management of heterogeneous data are achieved. Finally, correlation methods are applied in petroleum domain to provide support for effective management of petroleum scientific data and service sharing.

Web service is an emerging distributed computation mode and provides Internet-based software service unrelated to development environment. Such service-oriented calculation mode follows certain architecture, adapts Web environment openness and lays technological base for Web information and application procedure integration. But, existing Web service lacks obvious semantics. Web services cannot well understand information mutually conveyed, hinder service communication and interaction and cannot fulfill automatic service combination. From the perspective of service discovery, keyword-based grammar service discovery cannot ensure accuracy, lack semantic information support and cannot achieve secondary selection from given service candidate set. The key cause of the above problems is that there is lack of sharable semantic platform oriented to Web service. Thus, sharable semantic interpretation of Web service interface and execution process cannot be fulfilled. The setup of semantic platform contributes to improving service discovery quality and laying a foundation for automatic service combination. In allusion to these practical application demands, research status and existing problems, this paper mainly focuses on Web service semantic platform setup to propose Multi-technologies Supported Service-Oriented Architecture (MTSSOA). This architecture regards ontology as the core and take Web service process attribute description and Semantic coordination as assistive technology.