Why Read This
While information and knowledge management (I&KM) professionals occasionally demonstrate returns on their real-time data warehousing investments, most business intelligence (BI) architectures continue to rely on enterprise data warehouses (EDWs) as an aggregation point for historical data loaded in batch from operational repositories. Early adopters aside, I&KM professionals are increasingly rethinking their EDW architectures to optimize their infrastructures for real-time applications. Often they add real-time support, leveraging short batch windows, trickle-feed loading, changed data capture, and other ETL-acceleration approaches. Alternatively, I&KM pros may rely on architectural approaches like operational data stores, event stream processing, data federation, and information fabric, approaches that either supplement the EDW or bypass it altogether. These real-time approaches offer compelling strengths and perilous weaknesses, so I&KM professionals must do their homework if their goal is to design a more flexible EDW to support really urgent analytics.