Data Management
Options
Dr. Merle P. Martin
MIS Department
CSU Sacramento
Acknowledgments
Dr. Russell Ching (MIS Dept)
Source Materiel / Graphics
 Edie Schmidt (UMS) - Graphic Design
 Prentice Hall Publishing (Permissions)
 Martin, Analysis and Design of
Business Information Systems, 1995

Agenda
Why manage data?
 Definitions
 Typical problems
 Data Administrator
 The DBMS
 Distributing data

Why Manage Data?
Delayed output
(paycheck)
 Locate a resource
 Where is the stock
item stored?
 Where does the
employee work?

Why Manage Data?

Make resource decisions
 Should we turn account
over to collection agency?
 Should we send customer
letter asking why he / she
hasn’t shopped here in 6 months?
 Should we give employee overtime?
Why Manage Data?

Determine resource status
 Is there enough stock in
warehouse to satisfy this
customer’s order?
 How much should I order?
 What is the value of
this resource?
 balance sheet
Definitions

File: resource inventory:
 Material
 People
 Employees, customers
 Funds
 Customer balances
 Accounts Payable
Definitions

Data Organization
 Bit / byte
 Character
 Field
 Record
 File
 DBMS
Data Hierarchy for
Stereos to Go
Database
File
{
12345 Smith John A 123 Main Street Sacramento CA 95819
12345 Smith John A 123 Main Street Sacramento CA 95819
12345 Smith John A 123 Main Street Sacramento CA 95819
Record
Field
Character
(Byte)
Smith
10110011
Bit 1
Definitions
Views:
 Physical - how stored
 Logical - how viewed
and used
 Volatility: - % records that change
 Immediacy: rapidity of change

Storage Problems
Redundancy
 Accuracy
 Security
 Lack of data sharing
 Report inflexibility
 Inconsistent data definitions
 Too much data
 information overload

Data Administrator
Clean up data definitions
 Control shared data
 Manage distributed data
 Maintain data quality

Clean Up Definitions
Synonyms / aliases
 Standard data definitions
 names and formats
 Date of Birth (AJIS)
 mm/dd/yy (courts)
 dd/mm/yy (corrections)
 Data Dictionary
 COBOL

Control Shared Data
Local - used by one unit
 Shared - used by two
or more activities
 Impact of proposed program
changes on shared data
 Program-to-data element matrix
 Control or clearinghouse?

Manage Distributed
Data
Geographically dispersed
 whether shared data
or not
 Different levels of detail
 different management levels

Low
Infrequent
Frequency of Use
Very frequent
Quite old
Currency
Required Accuracy
Future
Time Horizon
High
Highly current
Historical
Aggregate
Wide
Scope
Well defined
Detailed Level of Aggregation
External
Operational
Control
Source
Management
Control
Internal
Strategic
Planning
Maintain Data Quality
Put owners in charge
of data
 verify data accuracy
and quality
 Fairbanks Court example
 Who owns the data?

Issue
Should the Data
Administrator control
ALL data,
or just that data that crosses
organizational boundaries?
WHAT DO YOU THINK?
The DBMS
Data Base Management
System: software that
permits a firm to:
 centralize data
 manage them efficiently
 provide access to applications
 such as payroll, inventory
DBMS Components
Data Design
Language (DDL)
 Data Manipulation
Language (DML)
 Inquiry Language (IQL)
 Teleprocessing Interface (TP)
 Martin, Figure 16-5

Designers
Teleprocess
DDL
Database
DML
Update
Applic. Software
Programmers
IQL
Interface
Retrieve
End-Users
IQL LANGUAGE
Data
Base
IQL
SELECT EMP-ID,
EMP-FIRSTNAME,
EMP-LASTNAME,
EMP-YTD-PAY
FROM
EMPLOYEE
WHERE EMPID=1234
.
3-level Database Model
James Martin
 Sprague / McNurlin,
Fig. 7-2, pg. 207

External Level (1)
User views (logical)
 By application program
 Each has unique view
 Schema / subschema

Schema and Subschemas
Physical Database
Individual
Views
Subschema
User
User
DBMS
DBMS Software
Schema
Overall View
of the Data
Subschema
User
User
Subschema
User
User
Enterprise Level (2)
Under control of Data
Administrator
 DBMS
 Implementation data removed
 passwords
 report views

Physical Level (3)
Schema
 Pointers
(e.g., next record)
 Flags
(e.g., record frozen)

Traditional Data
Models
Hierarchical - one parent
 Network
 more than one parent
 student to course, major
 Relational (tables)

Hierarchical Model
Project 1
Dept. A
Dept. B
Dept C
1
3
5
2
4
Employees
6
Network Model
John Smith
Jane Smith
Savings
Mortgage
Checking
Account Number
First Name Middle Initial
Last Name
...
Credit Limit
Customer
Order Number
Order Date
Account Number
Date Shipped
Orders
Order Number
Line Item Number Product Code Quantity
Line Items
Product Code Product Name
Price
Unit
Manufacturer Code
Products
Relational
Manufacturer Code Manufacturer Name
Manufac(turer)
Object-oriented DBMS
An object is:
 a piece of data PLUS
 procedures performed
on data PLUS
 attributes describing data
PLUS
 relationship between object
and other objects
Distributed Data

Goals:
 move processing as
close to users as possible
 allow several applications to run
simultaneously on same data
Distributed Types
Fragmented
 distribute data without
duplication
 users unaware of
where data located
 Segmented
 data duplicated
 one site has master file
 problem with data synchronization

Why Distribute?
Save money
 offload DB processes
to less expensive
machines (PCs)
 Lower telecommunications costs
 DB closer to users
 Decrease dependence on a single
computer manufacturer

Why Distribute
Move control closer
to owner
 Increased DBMS scope
 more varied types of data
 link at workstations
 Permit storage of multimedia data

True Distributed DB
Local autonomy
(ownership)
 No reliance on central site
 Continuous operations
 not affected by another site
 Data transparency
 Independence

Independence
Fragmentation
 Replication
 Hardware
 Software
 Networks
 Database

Problems With
Distributed Databases
Security
 Shared data
 simultaneous update
 Complexity
 Need telecommunications
infrastructure

Issue
Is data in your organization
totally distributed?
 How?
 Should it be?
 Why or why not?
Points to remember
Definition
 Typical problems
 Role of Data Administrator
 The DBMS
 Distributing data
