User:Modevs/sandbox

Design methods for business intelligence databases

Business intelligence (BI) databases are databases designed to maximize reporting and analytic value of data by aggregating and summarizing important pieces of information in such a way that large reporting queries can run more efficiently.

Business intelligence oriented databases differ from operational databases in that a BI database is designed to derive business information from existing data , which can be used to report on or gather meaning by aggregating existing data sources.

While operational databases are focused around data-input queries to accurately and in a normalized fashion manage data, business intelligence databases are focused on data-output, often at the expense of normalization and limiting redundancy. Since data often will not be updated however, this is preferable and allows for faster reporting queries.

Comparison of goals

Operations	Business Intelligence
Minimize data redundancy (normalization)	Optimize data to run quickly with a wide range of queries using mass amounts of data; de-normalized to provide quick retrieval of data.
Requires sub-second response time	Response time is important however as queries generally will take seconds, minutes or even hours optimization is aimed towards minimizing the impact of these large queries
Minimal derived data, derived data is typically generated on-demand	Large amounts of derived data to speed up report-based queries
Historical data often sacrificed for speed	Large amounts of historical data purposely kept to generate temporal reports