Corporate Information Factory
INTRODUCTION Inmon (2001) has stated that the Corporate Information Factory (CIF) consists of the following components: „X Data warehouse (DW) „X Data marts „X Data mining „X Decision support system application (DSS) „X Primary storage management „X Metadata management ¡K¡Ketc Not all the CIF components will be examined in this paper; however, this paper seeks to emphasis on the discussion of data warehouse and data mart as described in the following three sections: „X Section I - Identify the difference between the data warehouse and data mart „X Section II - Identify the issues for implementing the real-time data warehouse „X Section III - Discuss how to evaluate the data warehouse and data mart SECTION I - THE DATA WAREHOUSE VERSUS THE DATA MART 1. ... Also, it integrates the data from multiple sources, and transforming these data into meaningful information that allows the managers to perform useful and consistent business analysis. ... The data mart contains only a small amount of historical information and it only suits the needs of the department. Data mart can be standalone or linked to corporate data warehouse. There are two kinds of data marts, which are dependent (linked to corporate DW) and independent (i. ... The data source of a dependent data mart is fed by a corporate data warehouse. ... 3 The different between Data Warehouse and Data Mart The following table provides a comparative overview of DW and data mart: Data Warehouse Data Mart Data type and usage - Contains the most detailed data that is found in the corporation - Incorporates information about many subjects often the entire enterprise - Huge amount of historical data - Contains aggregated or highly summarized data - Only a portion of an enterprises data, perhaps data related to a business unit or work group - Small amount of historical data Queries - Ad hoc queries - No particular purpose or requirements - Advantage: Flexibility of queries - Structured repetitive queries - Based on set of user requirements - Advantage: Speed of queries User - Explorer - A user does random ad hoc queries, generating unpredictable responses on large units of data - Designed to suit the collective needs of the entire corporation - Framer - A regular user undertakes standard, repetitive queries on small units of data that usually have predictable responses - Designed to suit the needs of a department Cost - High ($3,000,000 - $5,000,000) ** - Much lower (typically well under $500,000) ** Implementation time - Long (12-18 months) ** - Implementation time is less than 6 months ** Control - Controlled by the entire corporation - The right of data ownership is controlled in the department level Structure - Normalized structure - Simple and designed for unknown future usage - Star join structure - Designed for specifically known requirements ** Sources: adopted from Jane (1998) The issue is which type of systems should be building first, a data mart or DW? ... Take a view of the following downsides of data mart: - Data marts cannot understand all the data in the corporate view completely. ... - Lacking the consistent information, consider one data mart says the yearly revenue is $25 million. Another department states the yearly revenue is $11 million based on information from their data mart. ... - Produce a detailed development schedule SECTION II - IMPLEMENT A REAL-TIME DATA WAREHOUSE In order to remain competitive in today¡¦s economy, companies are no longer afford to make business decisions only based on week-old or month-old information. They need to access the updated information instantly, and use this information to make better intelligence business decisions faster, understand customers and market trends, for examples: - Loaded real-time data from financial markets to drive stock change alerts. ... - Recognizing abnormal sales trends for a particular product that occurs within current date instead of months, and use this information to alerts the companies to move the stock of a product to a particular location that might run out before that happens. ... The real-time DW is replenished in real-time and it is providing users with the most up-to-date information possible. Instead of using ETL tools, a real-time DW uses advanced solutions that capture, transform and flow (CTF) data between multiple operational systems and DW in real-time for data replenishment while transforming the data into meaningful information.