Transcript Chapter 1: Introduction

Oracle 8i Data Warehousing
(chapter 1, 2)
Data Warehousing Lab.
석사1학기 HyunSuk Jung
Chapter 1 - Warehouse: What is it,
Who needs it, and Why?
 This chapter will help you what users require from a business
intelligence system(BIS) and why a data warehouse is often necessary
to satisfy these demands.
 We will answer the following questions.
2

What is business intelligence?

What are the business and technical goals of business intelligence?

What is data warehousing?

What are the business drivers of data warehousing?

What are the technical drivers of data warehousing?
DW
Data Warehousing
Lab.
Problems with the Current
Reporting Architecture
 Accessibility
: Can I get to my information when I need it?
 Timeliness
: How long after transactions occur do I get my information?
 Format
: What kind of reports can I get?
 Integrity
: Can I believe the data I get? Is it accurate?
3
DW
Data Warehousing
Lab.
The Goal: Business Intelligence
 The real goal of reporting systems is decision support – business
intelligence
 Business intelligence system is a system that give users access to
their data and allows them to analyze and format the data as
needed.
4
DW
Data Warehousing
Lab.
An automatic Teller Machine(ATM)For Data
 Figure 1.2 shows how IS has gone from being the conduit to
being the builder of the conduit
5
DW
Data Warehousing
Lab.
So, What’s Data Warehouse?
 Inmon describes the warehouse as
“subject-oriented, integrated, nonvolatile, time-variant
collection of data in support of management decisions.”
6
DW
Data Warehousing
Lab.
Subject Oriented
 Subject-oriented information is key in situations like this.(as manager)
 How could they allocate sales resources?
 How could they make production plans?
 How could they justify their huge bonuses?
7
DW
Data Warehousing
Lab.
Integrated
 Figure 1.5 illustrates this difficulty – the marriage of different coding
schemes into one for the warehouse data.
8
DW
Data Warehousing
Lab.
Nonvolatile
 Warehouse is read-only. Users can’t write back.
 Figure 1.6 illustrates this fundamental difference between OLTP and the
data warehouse.
9
DW
Data Warehousing
Lab.
Time Variant
 Time is a very important component of reporting and, thus, of data
warehouses.
10
DW
Data Warehousing
Lab.
Business Intelligence Differs from
Transaction Processing
 Difference between business intelligence computing and
operational computing
Operational system
Business intelligence system
Size
Small pieces of information
Huge blocks of information
Update
Frequently be updated in real
time
Almost never needs to be updated in
real time
Data input Rapid input
Enter no data, nonvolatile
Response Need immediate response
time
Don’t need such blazing response
times
Usage
pattern
Predictable
Not stable
Database
design
complex
Easy for users
11
DW
Data Warehousing
Lab.
Why Oracle8i for Data
Warehousing?
 Oracle8i – Relational Database
 Oracle Reports & PL/SQL – Development Tools
 Oracle Warehouse Builder(OWB) – ETL
 Express – Multidimensional Database Engine
 Discoverer – Relational OLAP Query Tool
 Oracle Data Mining Suite
12
DW
Data Warehousing
Lab.
Chapter 2 – Things to Consider
 Be Pragmatic
 Articles and Books contain Opinions, Not Facts
 Buyer Beware!
 Start with business Requirements – Not Technology
: data warehousing is not about technology. data warehousing is about solving
business problems
13
DW
Data Warehousing
Lab.
Data Mart or Data Warehouse?
 Data mart vs Data Warehouse

Data warehouse is a “broad” data store, contains a number of subject
areas.

Data mart focuses on a more narrow part of business, covers a single
subject area.
 Build small, but think big.

For example, the telecommunications department needs to analyze longdistance usage on a monthly basis. It will build its mart using two data
sources:
 It will gather data from the corporate data warehouse about every employee, their
phone number, and their departments.
 It will gather data from its long-distance provider about the long-distance usage
from each phone.
14
DW
Data Warehousing
Lab.
Data Warehouse Differences
 A Developer’s Perspective

Mindset
 data-capture
data distribution
 “How quickly can I insert this row in the database?”
“How can I deliver results to a query that summarizes 10 million rows in a
reasonable amount of time”

Denormalization
 Predicting the work that a user will request.
 “Predoing” it in a batch job so that the system doesn’t have to do it when the user
submits his or her query.
 The User’s Perspective
15

Unreliable data?

Poor response time?

Complex user interfaces?
DW
Data Warehousing
Lab.
Why Oracle for Data Warehousing?
 Total Solutions
 OLAP and Server Access
 Parallel Query Option(PQO)
: to provide for faster query throughput.
 Cross-Platform accessibility
: Oracle Open Gateway technology
 Data Acquisition and Transformation
: for moving and transforming data from source systems and it to the data
warehouse or data mart.
16
DW
Data Warehousing
Lab.