Data—the information that quantifies and describes your organization's activities—is the wellspring of any business insights that come from your BI environment. Developing a sound strategy to seek it out, clean it, marshal it, corral it, and move it into a useful posture is one of the most important steps in the entire process.
In order for users—by way of the software they use—to have easy access to your data, you'll need to place it into a target data environment. Once in this setting (which may take on a number of different forms), the data is available to BI user applications that can harvest and use it to produce the key strategic and operational insights that make the entire BI effort worthwhile.
The BI target database, no matter what form it takes, must be optimized for "data out" procedures—queries that feed reporting-and-analysis tools.
The target database must be built to withstand the pressures peculiar to BI usage. Normally the BI target database is geared to storing massive amounts of historical data, at all levels of summarization and aggregation. In addition, the BI target database may be designed to feed a diverse array of large, complex queries and reports—in addition to feeding high-end analysis tools.
This chapter discusses the most common BI target data environments: data warehouses, data marts, operational data stores, and hybrid models. As we go along, we consider these important questions: ...