A Data Warehouse Is Quite Different From Other Enterprise Systems
The most significant differences between the corporate data warehouse and other enterprise-wide systems involve the methods used to store and manage data.
A data warehouse has significantly different features from other enterprise-wide systems, particularly in how data is stored, managed and manipulated.
There are four key characteristics which separate the data warehouse from other major operational systems:
- Subject Orientation: Data organized by subject
- Integration: Consistency of defining parameters
- Non-volatility: Stable data storage medium
- Time-variance: Timeliness of data and access terms
Data Warehouse vs. Other Operational Systems
In the data warehouse, the subject-oriented database characteristic of the system organizes data according to subject, unlike the application-based database. The alignment around subject areas affects the design and implementation of the data found in the data warehouse.
This means that the major subject areas influence the most important parts of the key structure and the selection of the database software.
In many organizations, corporate data is often stored in several different forms and locations, and may come from mainframe applications or distributed systems and information-gathering devices throughout the organization.
Data entries also differ from applications-oriented data in the relationships among entries. Although operational data has relationships among tables based on the business rules that are in effect, the data warehouse encompasses a spectrum of time.
In addition, an operational data warehouse is an integrated system, because data is moved into it from many different applications, and this integration is noticeable in several ways:
- Implementation of consistent naming conventions
- Consistent measurement of variables
- Consistent encoding structures
- Consistent physical attributes of data
Inconsistent Data in the Data Warehouse
Operational data derived from other major systems is often inconsistent across applications. Pre-processing this information helps to reduce access time at the point of inquiry.
Data stored in a data warehouse is often five to ten years old, and is used for making consistent comparisons, viewing trends, and providing forecasting tools. It is accurate at any moment in time and will produce the same results every time for the same query.
On the other hand, operational environment data reflects only accurate values at the moment of access, although the data in such a system may change at a later point in time through updated data or inserts.
The time variant feature of the data warehouse is observed in different ways. In addition to the lengthier horizon when compared to the operational environment, time-variance is also present in the key structure of a data warehouse.
Every key structure contains implicitly or explicitly, an element of time, such as day, week, or month.
Time-variance is also evidenced by the fact that data in a traditional data storage facility is never updated, while operational data in a data warehouse is updated as the need arises.