A Comparative Analysis of Data Warehouse Design Methodologies for Enterprise Big Data and Analytics

Abstract

Data warehouses and business intelligence have become critical for modern enterprises looking to gain better insights and drive improved business outcomes with the business analytics based on data. With the exponential growth in enterprise data, many companies have turned to big data technologies and hybrid architectures for their data warehouses. Traditional data warehouse design methodologies must therefore be adapted or new methodologies and capabilities developed to meet the emerging needs of enterprise big data and analytics. This research article provides a comparative analysis of several leading data warehouse design methodologies - Inmon, Kimball, Data Vault, and Lambda Architecture - evaluating how they have evolved for enterprise big data use cases. The differences in core approaches, schema design, implementation architectures, and more are explored with a focus on enabling scalable, flexible and performant data warehouses for advanced analytics. Several example data warehouse architectures are presented across three comparative tables highlighting key differences between methodologies. Benefits, drawbacks and best use cases are covered for each methodology. The analysis shows that while all the methodologies have merits, the Lambda Architecture provides the most comprehensive approach combining best practices from the others for managing the complexities of enterprise big data.