Oracle Data Warehouse Topology
Download
Report
Transcript Oracle Data Warehouse Topology
State of Technology
Oracle
ISM 611
Dr. Hamid Nemati
11/8/1999
Francis Andoh-Baidoo
David S. Clark
Daniel B. Madeja
Phillip Planes
Kevin Thompson
Oracle Corporation
Overview
Founded in 1977
$9.1 billion in revenue for Q1 2000
Second largest application software company
Largest database software company
Operations in 145+ countries
Over 43,000 employees
Q1 Fiscal 2000
Revenue
$2.0 B
13%
Net Income
$237M
21%
EPS
$0.16 vs. $0.13
Pretax Margin
22.6% vs. 20.5%
Oracle Corporation
Q1 FY00 Revenue Highlights
Total
Revenue
Up 13%
License and Other Revenue
Up 9%
Geographic
Revenue
Growth
Americas
Up 12% (LC 14%)
Europe / Middle East / Africa
Up 9% (LC 13%)
Service Revenue
Up 16%
Asia Pacific
up 34% (LC 18%)
Revenue Distribution
Q1 Fiscal Year 2000
Product Revenues
68%
Consulting
and
Education
32%
Total Revenue $9.1 Billion
Geographic Distribution
Q1 Fiscal Year 2000
Asia Pacific
10%
EMEA
33%
Americas
59%
Oracle’s Two Businesses
Q1 FY00
Y-Y
Database
$6.6 B
19%
Applications
$2.5 B
25%
Total
$9.1 B
20 %
Oracle Leads 5:1 in UNIX ...
1998 RDBMS Market Share on Unix
Sybase
7%
NCR
4%
IBM
7%
Others
8%
Informix
13%
Oracle
61%
Source: Dataquest, March 1999
Database Growth
Opportunities
– E-business
– Data Warehousing
– Packaged Applications
– Release of 8i
– Middle Market
Nearly 2/3 of Fortune 100 Rely
on Oracle for E-Commerce
Oracle
64%
Source: Collaborative Research 1999
Others
36%
Data Warehouse
1997 WW Data Warehouse Data
Management Revenue by Vendor
Revenue($M)
600
500
400
Source: IDC, 1998
Teradata
SAS
IBM
0
Oracle
200
100
Microsoft
300
Oracle
IBM
Microsoft
SAS
Teradata
Oracle #1 in e-business
The only company with:
100% pure internet ERP applications
100% pure internet CRM applications
100% pure internet database
100% pure internet development tools
“If the Internet turns out not to be the
future of computing, we’re toast. But if it
is, we’re golden.”
Lawrence J. Ellison, Chairman and CEO of Oracle
Oracle & e-Business
New Model for Business
Based on the Internet
Not just e-Commerce
Components of e-Business:
– Internet Platform
– Enterprise-class Business Applications
– Services to Implement, Support and Manage
Oracle’s Internet Platform
Centered on Oracle 8i
Built-in Java Virtual Machine
Oracle Internet File System
Oracle interMedia
Development Tools
– Jdeveloper
– WebDB
Oracle Application Server
Oracle 8i Lite
Oracle Applications
Customer Relationship Management
ERP
Procurement
Self-Service Applications
Financial Applications
Human Resources
Strategic Enterprise Management Applications
Application Service Providers
Software is becoming a service
ISP’s + ISV’s = ASP
Oracle’s Business Online (world’s largest)
Benefits of ASP’s
–
–
–
–
–
No local infrastructure / resources required
Predictable Costs
Service Expertise
Automatic Upgrades
Data Back-Up
Oracle’s Service Organizations
• Oracle Consulting
• Oracle Education
• Oracle Support Services
Oracle Consulting
•
Leverages Oracle’s knowledge to develop individual
business solutions
•
Wide range of business solutions:
•
•
•
•
expand markets by identifying new customers,
creating new offerings, and extending the supply
chain
improve efficiencies by automating the value delivery
chain, lowering channel costs, and improving
productivity
identify and retain profitable customers, build brand
loyalty, and enhance customer service
design and implement a comprehensive data
warehousing solution
Oracle Education
•
World’s second-largest IT training provider, with
offices in more than 60 countries
•
Offered training services:
• instructor-led training
• interactive distance learning
• network-based training
• computer-based CD-ROM courses
• in-depth seminars
• customized classes
• training needs assessment
•
•
Certification programs
Organizational Change Management
Oracle Support Services (OSS)
•
Industry’s leading provider of support for
information management systems
•
More than 3,000 professionals support over
120,000 customers
• Support around-the-clock, seven days a week,
worldwide, and for any given product
•
Customers can choose between web-based
support, phone support, on-site support, and large
systems support
•
Different comprehensive support options are
available to meet individual requirements
Oracle’s MetaLink 2.0
•
•
•
A “change in Oracle’s support paradigm”
•
Provides end-users with personalized, flexible
self-service options
•
Allows end-users to log and update technical
assistance requests (TARs) online, receive
product updates and information tailored to their
individual needs, and download solutions
MetaLink 2.0 is an Internet Support Tool
Internet as a medium to deliver quality,
personalized and timely customer service
Oracle’s MetaLink 2.0 (cont.)
•
Empowers customers to prevent, analyze and
solve problems with the same information
sources used inside Oracle
• More than 180,000 customers use MetaLink
• Site is managed by more than 200 Oracle
engineers that provide timely posting and
updating of technical data
Source: Oracle, Darwin-Overview-WHT97.ppt
Oracle Warehouse
Comprehensive Set of Products
Oracle Business Intelligence
Reports
Operational
Data
Oracle 8i
ERP
Data
Warehouse
Builder
Application
Server
Discoverer
Darwin
External
Data
Oracle8i
CWM and Repository
Designer and Enterprise Manager
Express
Oracle8i
Internet Enabled
Scalable (with data and users)
Fast querying capability
Partitioning and “rolling windows”
Warehouse Builder 2.0
Fast implementation
Design, generation and loading
Written completely in Java
Graphical and wizard driven
Tightly integrated with Oracle8i
Common Warehouse
Metadata
Enables interoperability among vendor tools
Internet ready:
– Uses Java programming language
– XML for import/export
– UML as modeling language
Application Program Interface
Data Mart Suite 2.0
Data Mart Designer 2.1
Data Mart Builder 2.0
Oracle8 - Enterprise Edition
Enterprise Manager 1.5
Oracle Discoverer 3.1
Webb App Server
Reports and Reports Server 3.0
OLAP Tool:
Oracle Express Server
Distributed
over Web
Integrates
with DBMS
Big picture
and detail
Maintains
high levels of
data
integrity
Integration
with popular
tools
Scalability
Backed with
Oracle’s
experience
Analytical
functions
Oracle Business Intelligence
Darwin
Source: Oracle, Darwin-Overview-WHT97.ppt
Enterprise Manager
Business Intelligence Levels
Analytical
Tools Used
Top-Down
Methodology
Surface
SQL (Structured Query
Language) for simple
queries and reporting
Shallow
Statistical & OLAP
for summaries,
analysis, &
forecasting
Hidden
Data Mining
for classification,
clustering, and
predictions
Corporate Data
Bottom-Up
Methodology
Source: Oracle, Darwin-Overview-WHT97.ppt
Data Mining
Source: Oracle - Darwin Multimedia Presentation - www.oracle.com
Oracle
Darwin
June 1999 - To improve data
mining capability, Oracle
purchases Darwin from
Thinking Machines Corporation
Enterprise-wide Data Mining
Software - Darwin
Power and Ease of Use
Intuitive GUI
“Wizards” to guide
and automate
Powerful, scalable, parallel
UNIX server
MS Excel integration
Client: Windows
Servers:
– Sun Solaris
– HP-UX
– Windows NT (Release 4.0)
Source: Oracle, Darwin-Overview-WHT97.ppt
Data Management
Access data
– Data warehouses
– Relational databases
(ODBC)
– Support for SQL queries
– SAS files
– Flat files
Prepare data
–
–
–
–
sampling
randomization
computed fields
append, merge, select,
project, explode data
– replace and set form
Source: Oracle, Darwin-Overview-WHT97.ppt
Model Building
Induction Tree
Tree
+
- +
+
+
+
?
+ - +
+
- -
Classification & Regression Trees (C&RT)
– Classification and prediction of binary, multiclass, and continuous variables
– Decrease functions: gini and entropy
– Prune functions: cost and gini
Neural Networks
– Classification and prediction of binary, multiclass, and continuous variables
– Activation functions: sigmoid, hypertangent, and linear
– Training algorithms: conjugate gradient, modified Newton, steepest
descent, backpropagation, and genetic algorithm for optimizing neural networks
k-Nearest Neighbors (Memory-Based Reasoning)
– Trainable k-Nearest Neighbors approach
– Classification and prediction of binary, multiclass, and continuous variables
– User-definable weights for input fields
Neural Nets
Net
+
- +
+
+
? - +
- + + + - --
Clustering
– k-means
– Interactive graphics
Source: Oracle, Darwin-Overview-WHT97.ppt
Clustering
Cluster
? ?
? ?
?
?
? ??
? ?
?
k-Nearest Neighbors
Match
+
- +
++ +
?
- - +
+
+
- -
Enterprise-wide Data Mining
Software
Source: Oracle, Darwin-Overview-WHT97.ppt
Scalable, Parallel Performance and Model Accuracy
1
1
1
1
High
Darwin Module
Tree
Net
Match (train)
Number of Processors
2
4
8
1.93
3.80
6.30
1.92
3.36
7.37
2.00
4.04
7.97
Linear Scalability
Model Accuracy
7
6
5
Tree
4
Net
Match
(train)
3
2
Low
Performance Speed-Up
8
1
1
2
3
4
5
6
7
8
Number of CPUs running in parallel
9
Small
Data Size
Large
New Features in Darwin
Release 3.5
Data import wizard
– Databases
– Text
Missing value
treatment wizard
Model Seeker
Computed fields
Tree display
Key fields wizard
Multi-model
comparison
Lift charts for Tree
Editable workflow
More...
Source: Oracle, Darwin-Overview-WHT97.ppt
Cumulative Targets vs. Percent of Population
Cumulative Targets
Darwin Release 3.5 New
Features
Editable Workflow
Interactive means of querying workflow
objects
Ability to edit workflow
Source: Oracle, 35New-40preview97.ppt
New Features in Darwin
Release 3.6
Clustering
wizard
– k-means
ODBC
write-back
to the
database
Source: Oracle, Darwin-Overview-WHT97.ppt
New Features in Darwin
Release 4.0
Windows NT & UNIX servers
Native database access
Faster algorithms
Better parallelism
Enhanced clustering
Cumulative
Targets
Cumulative Targets vs. Percent of
Population
Source: Oracle, Darwin-Overview-WHT97.ppt
Benefits Summary
Understand customer
behavior
– Who is likely to churn?
– What are the profiles
of my customers?
– What is my expected
profitability?
– Where are my best retention
and cross-sell opportunities?
Mine more data, faster
Build better models
Generate business intelligence
Improve customer relationship management
Source: Oracle, Darwin-Overview-WHT97.ppt
Any
Questions
???