Talk to Anup: Difference between Bill Inmon and Ralph Kimball Da...: Difference between Bill Inmon and Ralph Kimball Data warehouse model Bill Inmon is first author to introduce Data warehouse design and...
Thursday 3 December 2015
Talk to Anup: Bill Inmon Data warehouse model
Talk to Anup: Bill Inmon Data warehouse model: Bill Inmon Data warehouse model and Development approach Bill Inmon is called Father of Data warehouse concepts, he is the first author ...
Talk to Anup: Ralph Kimball Data warehouse model
Talk to Anup: Ralph Kimball Data warehouse model: Ralph Kimball Data warehouse model and Development approach Ralph Kimball is well known author who introduced new pattern of data stor...
Wednesday 2 December 2015
Ralph Kimball Data warehouse model
Ralph Kimball Data warehouse model and Development approach
Ralph Kimball is well known author who introduced new
pattern of data storage in Data Warehouse, so called Dimensional Data
Architecture; Based on Fact and Dimension tables.
Kimball’s model is much more refined and simplified, where
all Subject oriented data you can store in Dimensions, and measures In fact
tables; whereas all dimension tables will be connected to fact tables by foreign
key relation.
Ralph Kimball introduced Star and Snowflake Schema, to store
data in Data Marts/ Data Warehouse.
Implementing this model is very easy and time saving. But Kimball’s
model mostly suitable for small and mid-cap industries like Automobile,
Education, and Retail.
Here is I have tried to Design Kimball’s DW model
Ralph Kimball’s Model follows Bottom to Top approach, that
is nothing but lowest level of detailed data is stored in Data Marts. That is the reason this
approach is called Bottom Up approach.
Here Subject Oriented data is stored in Dimension tables and Measures of same data will be stored in Fact tables.
Take an Example of
Students Enrollments, Students Information and Enrollment information; will be
stored in Dimension tables whereas number of Students enrolled per class, per
Year, percentage will be stored in Fact tables; And dimension tables will be
connected to fact table by foreign key relation.
To normalize data at maximum level Kimball introduced Snowflake
Schema. One dimension is further subdivided into another dimension, connected
with foreign key.
Please leave your comments, suggestions
Thank you!
Bill Inmon Data warehouse model
Bill Inmon Data warehouse model and Development approach
Bill Inmon is called Father of Data warehouse concepts, he
is the first author to introduce the Data Warehouse, a computer generated data
also can be stored as Data warehouse and use for further analysis and decision
making.
Bill Inmon mainly concentrated to store data as a centralized
repository. Where all different systems data will be stored including history
and Data Access layers can access unique information with history.
Inmon's definition says that Data warehouse is Subject
Oriented, Non- Volatile, Integrated and Time variant collection of Data.
I got opportunity to
work on both the design patterns; I tried to compare what exactly real-time
difference between these two design patterns.
Bill Inmon’s model mainly suits for big industry, large data
cap, like Insurance, Banking, and Finance where huge data collection will be
there, daily basis; Almost 1 to 10 TB data collection daily, now here is a big
question, how to manage this much of big data; Multiple transactions by single
user, lots of history by single users, one user holds multiple accounts, actually
this is a very big challenge for companies to manage the data, coming from
different systems.
Bill Inmon’s model gives robust solution for all challenges,
where you can have master tables or we can call it as a repository tables.
Repository Tables or Master Tables are partially or fully normalized (depends of requirements), holding the subject oriented master data; a
Unique record with history, so you can easily and exactly track latest updates
and previous transactions.
Once we have all Master tables as a Repository, The Subject
Oriented data Marts will be created from master data with reports specific Lookup
tables. Which we can directly use for Reports, Data Visualization, data mining,
Analytics, Predictive reporting.
All this is called Business Intelligent, It helps for decision making to the higher
management.
Please leave your comments, suggestions
Thank you!
Difference between Bill Inmon and Ralph Kimball Data warehouse model
Difference between Bill Inmon and Ralph
Kimball Data warehouse model
Bill Inmon is first author to introduce Data warehouse
design and Kimball has introduced new, relatively simpler Dimensional model
design, with Star and snowflake schema.
Here are some differences between Inmon and Kimball’s model.
Bill Inmon Model
|
Ralph Kimball Model
|
Inmon Model is little complex to design, and can have lot of
customization based on your own style.
|
Kimball Model relatively easier to develop, time saving and fixed
Star or Snowflake schema designs we can use.
|
Supports Huge industry needs.
|
Perfectly matches to medium and small company needs.
|
Time talking and long projects.
|
Comparatively less time taking.
|
Require Expertise engineers thought project.
|
Once developed, little experienced engineers can work.
|
Design Supports Enterprise applications and wide range of data.
|
Design supports individual and medium cap clients business area.
|
After development modification and customization is easier and can be
done faster.
|
After Development modification and customization is time taking
process.
|
Inmon Design uses ER model in Data warehouse.
|
Kimball Model uses Star and Snowflake Schema.
|
In Inmon’s design Data Marts are physically separated from Data warehouse.
|
Whereas in Kimball’s model it’s not necessary to separate Data marts
from Dimensional Data warehouse.
|
Inmon Design suggests First developing Central repository and then
developing Data marts from Central repository.
|
Whereas Kimball’s design suggest developing first critical Data Mart’s
which can supports all reporting and analytical needs.
|
Please leave your comments, suggestions.
Thank you!
Subscribe to:
Posts (Atom)