Tracing data warehouse design lifecycle semantically

Data warehouses (DW) are core parts of decision systems. Ontologies have largely contributed in designing DW systems, due to their ability to capture the precise semantics of design artifacts. Designing semantic DWs involves several steps, where each step is permanently evolving to satisfy new requirements offered by the technology progress. Passing from one phase to another requires important processes and decisions made by design actors. These decisions are usually lost once the DW is built. Tracing these decisions is a challenging issue. However, managing traceability in DW systems did not have the same spring as for software development. In this paper, we claim that the presence of ontologies can be an asset for managing DW traceability: ontologies can semantically define design artifacts during the whole design cycle, and their reasoning capabilities can be used to identify new trace links. In this paper, we propose an approach for semantic DW traceability that requires: (i) the formalization of each design phase, (ii) the identification of horizontal and vertical interactions (inside and between phases), (iii) their semantic definition, storage and usage. The approach is composed of three main activities for managing the traces. It is illustrated using LUBM benchmark, Protege Editor and Oracle semantic DBMS. It is implemented in a case tool assisting the designer for managing the DW traceability. HighlightsSemantic data warehouse (SDW) design is based on the availability of a domain ontology.Ontology (semantics and reasoning) is an important asset for defining a traceability strategy.The contributions include a traceability model and approach for SDW design.A case study proposed using Oracle semantic DBMS.Implementing prototype supporting the design process and the traceability approach.