From star schemas to big data: 20+ years of data warehouse research

Abstract

Data Warehouses are the core of the modern systems for decision making. They store integrated information extracted from various and heterogeneous data sources, making it available in multidimensional form for analyses aimed at improving the users’ knowledge of their business. Though the first use of the term dates back to the 80s, only during the late 90s data warehousing has emerged as a research area on its own, though in strict correlation with several other research topics as database integration, view materialization, data visualization, etc. This paper surveys more than 20 years of research on data warehouse systems, from their early relational implementations (still widely adopted in corporate environments), to the new architectures solicited by Business Intelligence 2.0 scenarios during the last decade, and up to the exciting challenges now posed by the integration with big data settings. The timeline of research is organized into three interrelated tracks: techniques, architectures, and methodologies.