What is Online Transaction Processing OLTP?
An OLTP environment is set up for a software application to facilitate the transactional nature of a business, doing such things as order processing, banking, stocks, and e-commerce applications.
A system optimized for OLTP will support a high volume of users running short-lived queries.
OLAP is structured to allow large sums of data to be analyzed, usually in an aggregated fashion, while also allowing the user to drill-down into the details. The system is designed in "layers" of related data, to allow the users to analyze the data by multiple dimensions.
A system optimized for OLAP, or decision support, will typically have few users with long-running queries. OLAP involves aggregation and multi-dimensional data, which can result in very large databases.
Today, OLAP has become an umbrella term for applications also known as decision-support (DSS), business intelligence (BI), or executive information systems (EIS).
MGI uses advance Power OLAP tools for accessing data in a multidimensional environment.
MGI's Power OLAP is set up for the end user rather than the IT department bringing more flexibility and less dependency on IT for reports, analysis, queries etc.
Data Mart vs. Data Warehouse - Fully Understood
A data mart is a departmental structure. Typical departments having a data mart are the finance department, sales department, marketing department and so forth. The data in the mart is designed for access optimization by the specific users of the data mart.
A data warehouse is structurally different from a data mart in that the data warehouse must serve the needs of the entire corporation. The data warehouse is truly a corporate structure, serving many different needs.
Typically the data mart is fed data from the data warehouse, whereas the data warehouse is fed data from the operational systems, old legacy systems as well as from other pockets of data such as flat files and spreadsheets.
|
|
Corporate |
Departmental |
Highly detailed |
Summarized / aggregated |
Normalized-efficient storage - non duplication of data. |
De-normalized, star joined design-less efficient storage but faster retrieval. |
Robust history |
Limited history |
Large volumes of data |
Limited volumes of data |
Data model driven |
Requirement driven |
Versatile |
Focused on departmental needs |
General purpose DBMS (database management system) |
Multi-dimensional DBMS technology |
|