Structuring Business Metadata in Data Warehouse Systems for Effective Business Support

Abstract

Large organizations today are being served by different types of data processing and informations systems, ranging from the operational (OLTP) systems, data warehouse systems, to data mining and business intelligence applications. It is important to create an integrated repository of what these systems contain and do in order to use them collectively and effectively. The repository contains metadata of source systems, data warehouse, and also the business metadata. Decision support and business analysis require extensive and in-depth understanding of business entities, tasks, rules and the environment. The purpose of business metadata is to provide this understanding. Realizing the importance of metadata, many standardization efforts has been initiated to define metadata models. In trying to define an integrated metadata and information systems for a banking application, we discover some important limitations or inadequacies of the business metadata proposals. They relate to providing an integrated and flexible inter-operability and navigation between metadata and data, and to the important issue of systematically handling temporal characteristics and evolution of the metadata itself. In this paper, we study the issue of structuring business metadata so that it can provide a context for business management and decision support when integrated with data warehousing. We define temporal object-oriented business metadata model, and relate it both to the technical metadata and the data warehouse. We also define ways of accessing and navigating metadata in conjunction with data.