 |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  | |
| By N2H | ||||||||||||||||||||||
Data Warehouse
August 4, 2008
If you're new here, you may want to subscribe to my RSS feed. Thanks for visiting and have a nice day!
Data Warehousing
Data warehousing represents an attempt to turn coporate data into a source of knowledge. A data warehouse has the principal aim of providing information that is truly useful for strategic decisions making. Information that constitutes a summarized and objective view of all areas of the enterprise.
A Data warehouse is a corporate decision support resource. It is a database that brings together data from diverse sources making it serve tactical/strategic decision needs.
One definition cites a warehouse as a support oriented, integrated, time variant and non-volatile collection of data in support of mangement’s decision making process.
Definition Quality of a Data Warehouse
Subject Oriented: i.e. data is organized along the lines of the key business entities e.g. suppliers, customers, stock e.t.c.
Integrated: the Data is from different subjects, sources and applications. It is held in such a way that it can all be applied collectively as one source
Qualities of Data
-
Time Variant; the data pertains to both current and past operations. The data occurs in layers parameterized by time.
-
Non-Volatile: the data is kept for a prolonged period of time without being changed or destroyed. However old data is stored in a more summarized form.
-
Specifically supports top level decision making.
-
Transaction levels are relatively low.
-
Applied at random intervals in a non-deterministic way.
Structure /Components of a Data Warehouse
-
Sources of operational Data
-
-
Old manufacture databases systems (corporate db systems)
Departmental Data (in databases or old filing systems)
External sources e.g. commercial data firms, internet etc
-
-
-
Load Manager (Front end component): this captures and prepares data, placing it in the data warehouses.
-
Warehouse Manager: This tool manages the data inside the data warehouse i.e. it does refining, back up e.t.c.
-
Query Manager (Back end component): Receives queries, schedules them and directs them to the correct data tables.
-
End user access tools: Used by the end user to place queries or otherwise process the data in a data ware house e.g. querry /reporting tools, data mining tools, online analytical processing tools.
-
Current data i.e. Data in the warehouse that pertains to recent operations. This is usually detailed data.
-
Summarized data: Data that pertains to past operations, retained in lightly or highly summarized form.
-
Archive data: old data images of the warehouse retained fort archive purposes.
Advantages of Data Warehousing
-
Very useful for trend analysis (assessing progress)
-
Improved quality of senior management decisions: good decisions often translate into better strategic advantage.
-
Great potential for competitive advantage in that analysis of the data warehouse may reveal patterns or trends that were previously unknown thus managers can capitalize on this information for competitive posturing.
-
High returns on investment
-
Improved managerial control because the manager has a broader and deeper weight into all areas of the enterprise which enables them to control the enterprise more effectively.
Disadvantages of Data Warehousing
-
Projects have a long time frame: the longer a project takes the more risky it becomes.
-
Some of the required data may be missing from operational systems; the data gaps presents a design challenge.
-
Other problems in the operational systems: if there is a problem with the system or operational system this problem may be transferred to the warehouse.
-
Data warehouse systems are high maintenance systems.
-
Underestimation of the data loading needs.
-
Need for large storage facilities: storage needs may be in the order of multiple petabytes.
-
Increased end user demands: When the warehouse is installed user requests for assistance and /or information may increase instead of reducing.
-
Data ownership; the data in the warehouse is a trully shared resource thus lack of data ownership.
-
Data homogenization: The data in the warehouse may infact not serve the intended purpose.
Data Marts
This is a small-scale data warehouse, specifically intended for departmental/divisional use.
The data mart may be a logical unit of the corporate data warehouse or it may be an independent resource.
The emergence of Data marts is a response to the challenges associated with the design and operation of enterprise wide data warehouses.
A Data mart is smaller and thus easier, cheaper and quicker to develop.
A data mart enables us to enjoy the benefits of a data warehouse, but on a smaller scale.
Comments
Got something to say?
You must be logged in to post a comment.
Powered by FireStats