Goal-ontology approach for modeling and designing ETL processes

Abstract

Abstract The data warehouse (DW) system’s design involves several tasks such as defining DW requirements, modeling DW schemas, and specifying ETL processes operations, and these have been studied and practiced for many years. However, the common designrelated problems such as defining user requirements and deriving data integration and transformation activities are far away from being resolved due to the variation and ambiguity of user requirements and the complexity of ETL processes, and these were the fundamental issues of conflicts in heterogeneous information sharing environments. Current approaches are based on existing software requirement methods that have limitations on reconciliation of the user semantics toward the modeling of the DW. This will prolong the process to define the ETL processes specifications accordingly. The semantic framework of DW system development in this paper is focused on the requirement analysis method for designing the ETL processes. The method RAMEPs (Requirement Analysis Method for ETL Processes) was developed to support the design of ETL processes by analyzing and producing the DW requirements as required by the organization, decision-makers, and developers. The ETL processes are modeled and designed by capturing two important facts: i) DW schemas, and ii) data sources integration and transformation. The validation process emphasized on the correctness of the goal-oriented and ontology requirement model, and validated by using compliant tools that can build both models. The correctness of RAMEPs was evaluated in the real case study of a gas utility company in Malaysia. The case study was used to illustrate how the RAMEPs method can be implemented for designing the ETL processes. The RAMEPs approach was reviewed by DW developers for assessing the strength and weakness of this method.