Design Life-Cycle-Driven Approach for Data Warehouse Systems Configurability

Abstract

Many modern software systems are designed to be highly configurable. Configurability is the ability to build consistent systems from a common architecture through selecting and synthesizing provided design elements. Configurability offers high customizability and efficient reuse strategy. Configurability has not enjoyed the same popularity in data warehouse (DW) design comparing to other types of software. Nowadays, we are assisting to an explosion of new DW applications due to high-performance computing and emerging hardware. This continuous evolution context reveals a high degree of variability that needs to be managed and exploited. We propose in this paper a configurability-aware approach for DW design, which allows designers to specify requirements defining suitable design options to generate a customized DW. To satisfy this objective, we need to perform the following three tasks: (i) a deep understanding of the DW design life-cycle analyzed by reviewing its evolutions, (ii) a formalization of each design phase and (iii) an identification of the interactions between phases. This analysis contributes in defining our approach containing: the configuration model which tailors the DW system to meet designers’ requirements and the configuration process which produces the corresponding DW configuration. The approach is defined using the Base, Variability, Resolution (BVR) models defined using the Common Variability Language proposed by the Object Management Group for defining variability modeling and implemented using BVR Tool. A case study providing two DW configurations is proposed to show the effectiveness of our approach.