Akinwale O. Falodun

1
Data Warehouse and Data Marts

• Akinwale O. Falodun
• CMSN 601
• Spring 2001

2
Terms to remember

• DW Data Warehouse
• DM Data mart
• Bill William

3
3
Agenda

• Define Data Warehouse and Data Mart
• Characteristics of DW and DM
• Contrast DW and DM
• DW, DM Architecture
• Why does a company need a DW
• Why does a company need a DM
• The debate - Ralph Kimball and William Inmon
• Problems with DW and DM
• Future of DW and DM
• Conclusion

3
4
What is Data warehouse?

• According to Bill Inmon, who is often called The
  Father of Data Warehousing, data warehouse is a
  subject oriented, integrated, time variant,
  non-volatile collection of data that supports the
  decision-making capabilities of the business
  user.
• Oracle Corporation defines data warehouse as a
  strategic collection of all types of data in
  support of the decision-making process at all
  levels of an enterprise.
• Important Definition Elements
• Subject Oriented, Integrated,Time Variant
• Non volatile collection of data, Decision-making
  capabilities

4
5
Characteristics of DW

• Central Repository
• Subject Oriented
• Integrated
• Time Variant
• Non volatile collection of data
• Viewing a Snapshot
• Consistent Data
• Decision-making capabilities

6
6
What is Data Mart?

• A data mart is an implementation of data
  partitioning strategy that enables targeted users
  to access functionally departmentalized data. A
  data mart is a subset of a data warehouse.
• It is also known as a business area warehouse or
  a departmental warehouse.
• Important Definition Element
• Data partitioning
• Targeted users
• Functionally departmentalized data
• Subset of a DW

5
7
Two Types of Data Mart?

• Dependent DM Fed by DW
• Independent DM Not fed by DW

5
8
Characteristics of DM

• Departmental Repository
• Departmentally Subject Oriented
• Time Variant
• Non volatile collection of data
• Consistent Data
• Decision-making capabilities

6
9
Critical Success Factor

• Clearly define scope of DW or DM
• Scope must not be changed during development
• Flexible design to accommodate future changes
• Identify Sponsors and Users
• Gain support for Requirements, Analysis,
  Prototyping and Testing
• Strategies should include Prototyping and
  Piloting
• Implementation and Deployment - one stage at a
  time
• Implementation stages should be meaningful and
  logical
• Software used should meet todays needs plus
  future

6
10
Hardware Software requirement

• Warning! - Please Address this at the early stage
• Server Products
• Express Server
• Web Server
• Processing and Modeling tools
• SQLLoader
• Gateways
• Queries and analysis
• Discoverer
• Express Analyzer

6
11
A Typical Data Warehouse Architecture
6
12
A Complex Data Warehouse Architecture
6
13
Contrasting DW and DM
14
William Inmon Vs Ralph Kimball
The data warehouse is nothing more than the
union of all the data marts - Ralph
Kimball Data Mart Does Not Equal Data Warehouse
- William Inmon

• KIMBALL
• Collection of DM DW
• DM grows big enough DW
• Dimensional modeling
• Build what the users need

• INMON
• DM extracted from DW
• DM growth big DM
• DM lacks thorough architecture
• DM not scalable
• DM lacks integration

15
Benefits of DW

• Competitive advantages
• Quick reaction to changes
• Adapting to changing customer needs
• access to data in different formats
• Rapid payback on investment within a short period
  of time
• Powerful strategic decision making tool
• Stable data, Subject Oriented
• Short term goal attainment, Long term goal
  assurance
• Hardware used is cost effective, robust
• Information management system

16
Benefits of DM

• In addition sharing some of the DW benefits...
• Departmental data
• Ease of access
• Ease of navigation
• Reduces Traffic into DW

17
Problems facing DW

• Overall Cost
• Expensive Hardware
• Expensive Sophisticated Software
• Expensive Skilled Team
• Training cost
• Time Consuming

18
Problems facing DM - Independent DM

• Redundant Data
• Redundant Processing
• Lack of Scalability
• Non-Integrated

19
Future of DW and DM
20
Conclusion

• You can catch all the minnows in the ocean and
  stack them together and still do not make a
  whale. Bill Inmon, January 8, 1998.

7
21

• End of presentation
• For informative articles, visit
• http//www.datawarehouse.com/iknowledge/articles
• QUESTION ?