International Journal of Informatics and Applied Mathematics

Yazarlar: ["Ali KHEBIZI", "Mourad HADJERİS"]

Konular:-

Anahtar Kelimeler:Business processes,Business Process Management,Data warehouse,ETL,Data integration,Big Data.

Özet: The large proliferation of the BPM discipline has considerably propelled companies' information systems which become ubiquitous. Consequently, a huge volume of data is generated during BP execution and stored in adequate IT systems. However, such data are often heterogeneous and the problem of their integration is posed with acuity in order to allow data analysis and to coordinate companies' activities. In fact, different techniques and approaches have been suggested in the research literature to tackle such issue. In the business intelligence specific context, this challenge is addressed by deploying the conventional ETL software tools allowing to build an unique container supporing various data and which is perceived as a data warehouse. Nevertheless, current distributed information systems are spanning enterprises boundaries and deal with a large variety of data sources having various formats and produced in a continuous manner. Thus, actual ETL tools seems unsuitable to handle massive data and they remain limited for facing to issues imposed by related to business processes execution data stored in log files. In this paper we propose an enhancement of the structure and functionalities of standard ETL tools in order to handle heterogeneous data integration generated by business processes execution. The improved system, named OLE-ST, constitutes a fundamental enrichment of the existing ETL mechanism. The proposed approach has been implemented in a software tool that ensures best performances for exploiting the target data warehouse to be built.

---