Traceability of Tightly Coupled Phases of Semantic Data Warehouse Design

Abstract

Ontologies have largely contributed in designing and exploiting decision systems. This is due to their ability to capture the semantics of design artifacts. Note that the design process of Data warehouses ($\mathcal {DW}$) projects involves several important and tightly coupled steps, where each step is permanently changing and evolving to satisfy the new requirements offered by the technology progress. Passing from one phase to another requires important and heavy processes and decisions made by design actors. These decisions are usually lost once the warehouse is built. Tracing these decisions is a challenging issue. In this paper, we propose to study the tractability management in the context of semantic data warehouses. We claim that the presence of ontologies can be an asset for the traceability, since the ontology can semantically define the design artifacts and their transformations during the whole cycle. To do so, we propose an approach for semantic data warehouse tractability that requires: (i) the formalization of each design phase and (ii) the identification and storage of horizontal interactions (inside the phase) and vertical interactions (between phases). The approach is illustrated using LUBM benchmark. It is implemented in a case tool assisting the designer for managing the $\mathcal {DW}$ traceability.