Storage Frameworks for Large Models within Model-Driven Data Warehouse Metadata Management Systems: Criteria and Evaluation

Abstract

Many metadata arise during the process of data warehouse engineering (DWE). In order to achieve a maintainable data warehouse (DW), this metadata should be organized in a metadata management system (MDMS). Modern software development technology suggests a model driven approach. Following this approach, the emerging metadata are stored in form of models and metamodels. The DW domain usually comprises large models with strong dependencies between each other. This characteristic has to be considered when building performant MDMSs. This paper defines a set of criteria, which storage frameworks for large models within model-driven MDMSs should meet to ensure a performant and secure business intelligence (BI) system. Furthermore, it presents an evaluation, on the basis of the defined criteria set, of latest storage frameworks based on EMF/Ecore technology.

Author-supplied keywords

Cite this document ^(BETA)