IAS Inc

Data Warehouse:
The Choice of
Inmon versus Kimball
Ian Abramson
IAS Inc.

Agenda

The 2 Approaches
D Bill

Inmon Enterprise Warehouse (CIF)

D Ralph Kimball Dimensional Design

Similarities
Differences
Choices

DW History

1990
D

1996
D

Inmon publishes Building the Data Warehouse

Kimball publishes The Data Warehouse Toolkit

2002
D

Inmon updates book and defines architecture for collection of

disparate sources into detailed, time variant data store.
The top down approach

Kimball updates book and defines multiple databases called data

marts that are organized by business processes, but use
enterprise standard data bus
The bottom-up approach

The Data Warehouse Is:

Bill Inmon, an early and influential practitioner, has formally defined a

data warehouse in the following terms;
D
D
D
D

Subject-oriented

The data in the database is organized so that all the data elements relating to the
same real-world event or object are linked together;

Time-variant

The changes to the data in the database are tracked and recorded so that reports
can be produced showing changes over time;

Non-volatile

Data in the database is never over-written or deleted - once committed, the data
is static, read-only, but retained for future reporting; and

Integrated

The database contains data from most or all of an organization's operational

applications, and that this data is made consistent

Ralph Kimball, a leading proponent of the dimensional approach

to building data warehouses, provides a succinct definition for a
data warehouse:
D

A copy of transaction data specifically structured for query and analysis.

Ref: wikipedia

I AS Inc

What are they saying?

These two influential data warehousing experts represent

the current prevailing views on data warehousing.

Kimball, in 1997, stated that

D
D

Inmon responded in 1998 by saying,

D
D

"...the data warehouse is nothing more than the union of all the data
marts",
Kimball indicates a bottom-up data warehousing methodology in which
individual data marts providing thin views into the organizational data
could be created and later combined into a larger all-encompassing
data warehouse.
"You can catch all the minnows in the ocean and stack them together
and they still do not make a whale,"
This indicates the opposing view that the data warehouse should be
designed from the top-down to include all corporate data. In this
methodology, data marts are created only after the complete data
warehouse has been created.

IAS Inc

What is a Data Warehouse:

The single organizational repository of
enterprise wide data across many or
all lines of business and subject areas.
D Contains

massive and integrated data

D Represents the complete organizational view
of information needed to run and understand
the business

IAS Inc

What is a Data Mart?

The specific, subject oriented, or departmental
view of information from the organization.
Generally these are built to satisfy user
requirements for information
Multiple data marts for one organization
D A data mart is built using dimensional modeling
D More focused
D Generally smaller, selected facts and dimensions
D Integrated
D

I AS Inc

Data Warehouses vs.

Data Marts
Data

Warehouses:

DScope

independent
Centralized or Enterprise
Planned
DData

Specific

application
Decentralized by group
Organic but may be planned
DData

Historical,

detailed, summary
Some denormalization
DSubjects

Some

history, detailed, summary

High denormalization
DSubjects

Multiple

subjects

DSource

Single

central subject area

DSource

internal and external sources

DOther

Few

internal and external sources

DOther

Flexible

Restrictive

Data

Project

oriented
Long life
Single complex structure

Marts:

DScope

Application

Many

Data

oriented
Short life
Multiple simple structures that may
form a complex structure

I AS Inc

The Inmon Model

Consists of all databases and information

systems in an organization..
D The

CIF (Corporate Information Factory)

Defines overall database environment as:

D Operational
D Atomic

data warehouse
D Departmental
D Individual

The Warehouse is part of the bigger whole (CIF)

IAS Inc

The Data Warehouse

Operational
(Day-to-Day Operations)
* Transactions *
Atomic Data Warehouse
(Data manipulated & moved)
* Transactions *
Departmental
(Focused)
* Source is ADW *
Individual
(Ad hoc)
* Source is ADW *

Customer Credit Rating

Customer Credit History

Customer by Postal Code

Delinquent Customers

IAS Inc

Inmon Modeling
Three levels of data modeling
D ERD

(Entity Relationship Diagram)

Refines entities, attributes and relationships

D Mid-Level

model (*DIS*)

Data Item Sets

Data sets by department
Four constructs:
D
D
D
D

Primary data groupings

Secondary data groupings
Connectors
Type of data

D Physical

data model

Optimize for performance (de-normalize)

I AS Inc

Relationship between Levels One and Two

of Inmon's Data model (Inmon,2002)

I AS Inc

The Warehouse Architecture

I AS Inc

The Inmon Warehouse

Data Sources

Staging

The Data Warehouse

Data Access

Source
DB 1

Source
DB 2

Landing
Staging Area

Data Marts

File or
External Data

Cubes

14

I AS Inc

The Kimball Approach

The Dimensional Data Model

D Starts

with tables

Facts
Dimensions

D Facts

contain metrics
D Dimensions contain attributes

May contain repeating groups

Does not adhere to normalization theory

D User accessible
D

I AS Inc

The Kimball Data Lifecycle

Data
Sources

Staging

The Data
Warehouse

Data Access

Source
DB 1

Workstation Group

Source
DB 2

Landing
Staging Area

End Users

File or
External Data

Cubes

16

IAS Inc

The Dimensional Model

IAS Inc

The Kimball Data Bus

Data is moved to staging area
D

Data is scrubbed and made consistent

From Staging Data Marts are created

Data Marts are based on a single process
Sum of the data marts can constitute an
Enterprise Data Warehouse
Conformed dimensions are the key to success

IAS Inc

The Kimball Design Approach

Select business process

Declare the grain
Choose dimensions
Identify facts (metrics)

IAS Inc

Kimballs Philosophy
Make data easily accessible
Present the organizations information
consistently
Be adaptive and resilient to change
Protect information
Service as the foundation for
improved decision making.

IAS Inc

Getting Started with Choices

Kimball
D Will

start with data marts

D Focused on quick delivery to users

Inmon
D Will

focus on the enterprise

D Organizational focus

I AS Inc

Kimball vs. Inmon

Inmon:
D Subject-Oriented
D Integrated
D Non-Volatile
D Time-Variant
D Top-Down
D Integration Achieved via an Assumed Enterprise
D Characterizes Data marts as Aggregates

Kimball
D Business-Process-Oriented
D Bottom-Up and Evolutionary
D Stresses Dimensional Model, Not E-R
D Integration Achieved via Conformed Dimensions
D Star Schemas Enforce Query Semantics

22

Data Model

I AS Inc

The Comparison
(Methodology and Architecture)
Inmon

Kimball

Overall approach

Top-down

Bottom-up

Architectural structure

Enterprise-wide DW
feeds departmental DBs

Data marts model a

business process;
enterprise is achieved
with conformed dims

Complexity of method

Quite complex

Fairly simple

Reference: http://www.bi-bestpractices.com/view-articles/4768

I AS Inc

The Comparison
(Data Modeling)
Inmon

Kimball

Data orientation

Subject or data driven

Process oriented

Tools

Traditional (ERDs and

DIS)

Dimensional modeling;
departs from traditional
relational modeling

End user accessibility

Low

High

Reference: http://www.bi-bestpractices.com/view-articles/4768

I AS Inc

The Comparison
(Dimensions)
Inmon

Kimball

Timeframe

Continuous & Discrete

Slowly Changing

Methods

Timestamps

Dimension keys

Reference: http://www.bi-bestpractices.com/view-articles/4768

I AS Inc

Inmon Continuous & Discrete

Dimension Management

Define data management via dates in

your data
D Continuous

time

When is a record active

Start and end dates
D Discrete

time

A point in time
Snapshot

I AS Inc

Kimball Slowly Changing

Dimension Management
Define data management via versioning
D Type

Change record as required

No History

D Type

II

Manage all changes

History is recorded

D Type

III

Some history is parallel

Limit to defined history

I AS Inc

The Comparison
(Philosophy)
Inmon

Kimball

Primary Audience

IT

End Users

Place in the
Organization

Integral part of the

Corporate Information
Factory (CIF)

Transformer and retainer

of operational data

Objective

Deliver a sound technical

solution based on proven
methods

Deliver a solution that

makes it easy for end
users to directly query
data and still have
reasonable response rate

Reference: http://www.bi-bestpractices.com/view-articles/4768

IAS Inc

How to Choose?
Characteristic Favors Kimball Favors Inmon
Nature of the
organization's
decision support
requirements

Tactical

Strategic

Data integration
requirements

Individual business
areas

Enterprise-wide integration

Structure of data

Business metrics,
performance
measures, and
scorecards

Non-metric data and for data that

will be applied to meet multiple
and varied information needs

Scalability

Need to adapt to
Growing scope and changing
highly volatile needs requirements are critical
within a limited scope

IAS Inc

How to Choose?
Characteristic Favors Kimball Favors Inmon
Persistency of data

Source systems are

relatively stable

High rate of change from source

systems

Staffing and skills

requirements

Small teams of
generalists

Larger team(s) of specialists

Time to delivery

Need for the first

data warehouse
application is urgent

Organization's requirements allow

for longer start-up time

Cost to deploy

Lower start-up costs,

with each
subsequent project
costing about the
same

Higher start-up costs, with lower

subsequent project development
costs

IAS Inc

References

Data Warehousing Battle of the Giants: Comparing the Basics of the

Kimball and Inmon Models: by Mary Breslin http://www.bi-

bestpractices.com/view-articles/4768

Inmon CIF glossary: http://www.inmoncif.com/library/glossary/#D

The Data Warehouse Toolkit, Kimball, 2002

Inmon, W.H. Building the Data Warehouse (Third Edition),
New York: John Wiley & Sons, (2002).
Kimball, R. and M. Ross. The Data Warehouse Toolkit: The
Complete Guide to Dimensional Modeling (Second Edition),
New York: John Wiley & Sons, 2000.

IAS Inc

Thanks and Questions?

Ian Abramson
IAS Inc
416-407-2448
ian@abramson.ca