Becoming BI Wise: Choosing the way to corporate knowledge
Enterprises have to make a choice of their infrastructure for business intelligence before they can begin to do the mining. The options are a data warehouse, a data store or web services that tie the existing infrastructure. Above all, enterprises want to be able to make decisions as close to real time as possible. On the other hand, they have to separate their operational databases from analytical processes so that the former are not slowed down in the performance of their routine business processes. The data from the operational databases is extracted, transformed into analytical categories (or dimensions) and loaded into data warehouses. The enhancement of the data for analytical processes is necessarily time consuming and delays the process of decision making.
Two alternatives are available to companies in order to overcome the delays in decision making when data is stored in data warehouses. Companies can choose to tie their existing operational databases, with web services or middleware, which links their disparate databases into a single network. The data is fed directly from the operational databases to the business intelligence application that does the analytical processing. Alternatively, companies can choose to create a data store or a server which is a mirror image of the data in their existing operational databases. A data store does hold as much historical information as does a data warehouse but the data latency is lower with it.
The chief advantage of a data warehouse is that operational data is transmuted into categories that are amenable for decision making. In an operational database, for example, information on sales would be available for each zip code. The data warehouse aggregates the data by individual geographical regions. An ETL engine in data warehouses joins the numerous tables such that the data is aggregated for individual regions; it categorizes the data by state, metropolitan service area, attributes purchases to a salesperson, etc. When data is accessed directly from operational databases into analytical packages, the time spent in making sense of the data would be high especially if the information is sourced from many different sources.
Without a doubt, the data warehouse is the technology of choice for data analysis and decision-making. Yet the industry continues to be frustrated by its high costs, batch processes in the transfer of data and long time required for queries. In recent times, some companies like Netezza have entered the industry with data warehouse appliances or special purpose data warehouses which offer less capability but better performance at lower costs.
Related Information
BI on a Budget
by Michael L. Gonzalez
Apr 17, 2004
The No-Sacrifice, Affordable Data Warehouse App
by Michael L. Gonzalez
Oct 30, 2004
Wrestling Value from All that Data
by Harriet Fryman
Mar 20, 2004
Applying All Your (Business) Intelligence
by Harriet Fryman
Mar 14, 2004
A Fresh Look at Data Warehouse Technology: Introducing the Data Warehouse Appliance
by William McKnight
Sponsored Links
5 Minute Guide to CRM Software | Business Intelligence Guide | Small Business CRM | Customer Relationship Management
|
