Transcript The Importance of IS Management

Managing Information Resources
Chapter 7
Information Systems Management In Practice 5E
McNurlin & Sprague
Introduction

Data vs. Information vs. Knowledge




Copyright 2002 by
Prentice Hall, Inc.
Data: facts
Information: data in context
Knowledge: information with direction or
intent - it facilitates a decision or action
Companies’ greatest asset: the
knowledge “embedded” in employees’
heads (tacit knowledge)
7-2
Managing Corporate Records: The
Problem: Inconsistent Data Definitions



Copyright 2002 by
Prentice Hall, Inc.
Problem: data definitions incompatible
from application to application, from
department to department, from site to
site, and from division to division
Reason: to get application systems up
and running quickly, system designers
sought data from the cheapest source
or politically expedient source
Result: different files with different
names for same data and same name
for different data
7-4
Managing Corporate Records: The
Problem: Inconsistent Data Definitions




Copyright 2002 by
Prentice Hall, Inc.
In order to get corporate data under
control, use of DBMS - database
management software
A “database administrator” to manage
DBMS to improve the problem of
inconsistent and redundant data
Broader definition of the data
administration role - Standard format
for data crossing organization
boundaries
Effective use of data dictionaries to
7-5
control standard data definitions
The Four Roles of Data
Administrators


Copyright 2002 by
Prentice Hall, Inc.
Clean up data definitions: getting rid of
data redundancies and inconsistencies,
e.g., two names for same data item,
design integrity flags, and train users on
meaning of data
Control shared data - data used by two
or more units - and analyze impact of
changes to programs that use shared
data; data dictionaries provides one
place to look for all uses of data.
7-6
The Four Roles of Data
Administrators


Copyright 2002 by
Prentice Hall, Inc.
Manage distributed data - geographically
dispersed - it also may cross hierarchical
levels of the organization
Maintain data quality: involves the
design and implementation of
procedures to maintain quality of data,
usually owners of data edit and verify
the accuracy, defining who owns the
data
7-7
The Importance of Data
Dictionaries



Systems and procedures for storing and
handling an organization’s data definitions
Purpose: eliminate errors of
understanding, ambiguities, and difficulties
in interpreting data
Ideal sequence:




Copyright 2002 by
Prentice Hall, Inc.
Set up the data administration function
Develop data standards
Purchase and install a DBMS
Install the data dictionary as the first dB
application
7-8
Case Example: Monsanto


International, decentralized.
Vision:





Copyright 2002 by
Prentice Hall, Inc.
responsive to customer needs
thinking and acting from global perspective
taking risks to enter new markets
treating earth as a closed system consumption and contamination cannot be
sustained
creating a thriving environment for
employees
7-9
Monsanto (cont.)

Three enterprise-wide IT projects:
 Financial Transaction System: SAP - covers all core
business transactions, including finance, order processing,
inventory management, product planning, and
manufacturing resource planning.
 Knowledge Management Architecture: Data warehouses
that can be sliced and diced with drill-down capability;
compare and leverage information.
 Enterprise Reference Data: SAP repository of master table
information in the company. Includes vendors, customers,
suppliers, etc. with different views for purchasing,
accounting, etc. Enables integration.
Copyright 2002 by
Prentice Hall, Inc.
7-10
Monsanto (cont.)

Getting the data into shape - ERD
Stewardship Dept.
Set data standards and enforce
quality
 Entity specialists, key managers. with
the greatest stake in the quality of
data
 Analysts who manage the data - a
global resource that the entire
company uses

Copyright 2002 by
Prentice Hall, Inc.
7-11
Managing Data: The Three-Level
Database Model
See Figure 7-4



Level 1 - The external, conceptual, or local level,
containing the various “user views” of the
corporate data that each application program uses
Level 2 - The logical or “enterprise data” level,
encompassing all an organization’s relevant data,
under the control of the data administration.
Level 3 - The physical or storage level, specifying
the way the data is physically stored
Copyright 2002 by
Prentice Hall, Inc.
7-12
The Three-Level Database Model:
Advantages



Copyright 2002 by
Prentice Hall, Inc.
Logical data can be separated from the
method of physical storage, different
physical devices can be used without
changing application programs.
Logical data relationships can vary from
different programs that use the data,
without requiring data redundancy.
Applications can use a subset of the DB
and organize it, again without
redundancy, in the best manner for the
application.
7-13
Four Data Models
1. Hierarchical model: structures data so
that each element is subordinate to
another in a strict hierarchical manner,
like boxes in organization chart
2. Network model: allows each data item
to have more than one parent,
relationships stated by pointers
Copyright 2002 by
Prentice Hall, Inc.
7-14
Four Data Models
3. Relational DBMS: where the
relationships among data items are not
expressly stated by pointers. Instead it
is up to the DBMS to find the related
items, based on the values of specified
data fields. Store data in tables, each
row=tuple, col.=attributive entity, DB
of choice today.
 See Figure 7-5 for relational operations
Copyright 2002 by
Prentice Hall, Inc.
7-15
Four Data Models
(cont.)




Designed to widen the use of DB to new kinds of
applications: CAD, medical applications,
knowledge representation for AI.
Retains traditional DBMS features and add two
new concepts.
 Object management: The mgt. of complex
kinds of data, e.g., multimedia and
procedures
 Knowledge Management: the management of
large numbers of complex rules for reasoning
Objects can be of any type: e.g., spreadsheet,
video clip, photograph.
A collection of objects is an object-base or
7-18
object-oriented database.
Copyright 2002 by
Prentice Hall, Inc.
A Look to the Future
 NASA needs to store 1016 bytes of satellite images,
enough to fill 10K optical disk jukeboxes
 CAD data for a skyscraper
 DNA sequence of human genome, several billion
elements long for genetic make-up research
 Customer buying patters at large retail chains - data
mining
 Multimedia DB of insurance policies
 Design databases should notify designers when
another designer has made a modification to the
system
 Need new data models to handle spatial data, time,
and uncertainty
7-19
Copyright 2002 by
Prentice Hall, Inc.
A Look to the Future



A single worldwide file system (?)
Distributed heterogeneous DB - for
inter-company DB
Security - authenticating enquirers
Copyright 2002 by
Prentice Hall, Inc.
7-20
Distributing Data


Distributed databases - 12 rules for a
distributed database, see Figure 7-6.
Operating principles depend on underlying
DB being relational.
SQL - Standard Query Language - a
standard language for accessing relational
DB.



Copyright 2002 by
Prentice Hall, Inc.
A data definition language for creating relational
tables, indexes to data, and fields of data
A data manipulation language for entering
information into a DB
A data control language for handling security
7-21
functions
Alternatives to “True” Distributed
Databases



Downloaded data files: e.g., sending data
from mainframes to PCs, most popular
method for distributing data
Copies of data stored at nodes: helps
process customer activity during the day,
night official updating of the master files
Not fully synchronized DB: second copies
in cache, errors cause a new primary
copy
Copyright 2002 by
Prentice Hall, Inc.
7-22
Alternatives to “True” Distributed
Databases


Server-based DB: In a distributed DB, each
node has a copy of the DBMS and
dictionary. In server DB, applications must
know where data is located; does not
support location transparency. Appropriate
for higher-performance transaction
processing.
Federated databases: Existing DB defined
independently and retain rules for others to
access its data. Works when incompatible
DB (text, alphanumeric, and image) are
needed in a single application.
7-23
Copyright 2002 by
Prentice Hall, Inc.
Managing Information: Issues
1. Value Issues: Information’s value depends on the
recipient and the context. The only practical way
to establish the value of information is to put a
price on it, and see if anyone buys. Tools:
 Information maps: Point to the location of
information, e.g., where to get quick answers
to questions
 Information guides: People who know where
information can be found.
 Business documents: Provide organization and
context - what documents an organization
needs
 Groupware: People to share information across
distances
7-24
Copyright 2002 by
Prentice Hall, Inc.
Managing Information: Issues
2. Usage Issues: Deals with how people use
information
 Information’s complexity needs to be
preserved - information should not be
simplified to be made to fit into a computer,
because this truncates sharing and
conversations.
 People do not easily share information, even
though its value grows as it is shared.
Culture often blocks sharing.
 Technology does not change culture.
Copyright 2002 by
Prentice Hall, Inc.
7-25
Managing Information: Issues
3. Sharing Issues:





Copyright 2002 by
Prentice Hall, Inc.
A sharing culture must be in place or the existing
disincentives will thwart the use of sharing
systems.
Information architectures have failed because they
do not take into account how people use the
information.
Sharing of corporate performance figures is
beneficial, but sharing of rumors can be demoralizing.
Separating information from non-information is an
information management issue.
Getting value out of information requires more
than a technology.
7-26
Four Types of Information

Fig. 7.2 presents a matrix representative
of these four types of information.



Copyright 2002 by
Prentice Hall, Inc.

Internal/record-based has been the focus of
attention of information systems, the type of
information computer-based applications
generate and manage easily.
External/record-based: can be accessed over
the Internet via public DBs.
Internal/document-based: intranet document
management 95% of the information. In most
organizations it is in document form, 5% data.
External/document-based: WWW
7-28
Four Types of Information

1
2
3
4
Copyright 2002 by
Prentice Hall, Inc.
Figure 7-8 lists the four categories of
information and shows typical
corporate authority, sources of
information, and examples of
technologies used in managing each:
Internal record-based
Internal document-based
External record-based
External document-based
7-29
Toward Managing Knowledge


Copyright 2002 by
Prentice Hall, Inc.
The more people are connected, and
the more they exchange ideas, the
more their knowledge spreads and can
thus be leveraged.
Process is the key - how do we transfer
tacit knowledge? IT is an enabler.
7-30
Conclusion

The job of the IS department is widening:




get data into shape
create and build an infrastructure for managing the full
range of information types
helping the firm leverage the tacit knowledge of its
employees
Companies that address all three areas,
and start implementing IT-based programs
in all three, will have a significant edge over
their competitors because they will be able
to leverage their intellectual assets.
Copyright 2002 by
Prentice Hall, Inc.
7-31