A decision support system is a software application designed to let an organization analyze data in order to support business decision making. You can use two different architectures to design decision support systems:
A centralized architecture contains definitions for the entire enterprise, providing a consolidated view
A set of databases, one for each business function or department, specify data using the same time and geographic dimensions, providing consistent comparisons
You can use Sybase IQ to implement either architecture, according to your analysis goals.
While specific requirements may vary, most databases used in decision support systems are specially adapted. Conventional relational databases used for running business processes are tuned for OLTP (On-Line Transaction Processing) and are not optimal for analysis, as shown in the following table:
To run process |
To analyze business |
|---|---|
Clerical users |
Managerial, clerical and analytical users |
Current data |
Historical data |
Detailed data |
Summarized and detailed data |
Highly changeable data |
Stable data |
Supports day-to-day operations |
Supports strategic decisions |
Transaction driven |
Analysis driven |
Optimized for structured queries |
Optimized for ad hoc queries |
When the database supports a specific functional department rather than being enterprise-wide, it is often referred to as a data mart. You can also think of a data mart as an application-specific database that focuses on a specific business problem.
Data marts can deliver the business intelligence required to gain competitive advantage at a modest cost and with exceptional ease of administration. Companies may have a number of business-oriented data marts supported by a central data model along with a central data staging and consolidation warehouse.
Decision support analysis has become less batch-oriented and more interactive. The business market demands rapid response to queries. With data accessible, more queries can be generated and decision making is improved.
