Towards a conceptualization of ETL and physical storage of semantic data warehouses as a service

摘要

The data warehouse technology has become the incontestable tool for businesses and organizations to make strategic decisions to ensure their competitively. The construction of a data warehouse (DW) passes by selecting relevant information sources, extracting relevant data and loading them into the DW. These processes require a precise expertise from designers related to logical and physical implementations of information sources, which is not usually an easy task. The diversity and heterogeneity of information sources makes the construction process of the DW complex and time consuming. Domain ontologies have been proposed to reduce heterogeneity between sources, platforms, services, etc. They resolve syntax and semantic conflicts. The phenomenon of adopting domain ontologies by organizations creates a new type of databases, called semantic databases (SDB). As a consequence, they become a candidate for building the semantic DW (SDW). To handle the diversity of information sources and hide the implementations aspects of sources, proposing a generic framework for constructingDW becomes a necessity. In this paper, we first proposed an ontology-based approach for designing SDB. Secondly, ETL phases are defined at ontological level to hide the implementation details. Thirdly, a storage service for ontologies and its associated data is given. Finally, our proposal is validated through a case study and a tool.