Download Database Design - KBU ComSci by

CS523 Database Design
Instructor : Somchai Thangsathityangkul
You can download lecture note at http://kbucomsci.weebly.com/
2
Class Presence
10%
Quiz
10%
Project
20%
Midterm
20%
Final
40%
Total
100%
Database Design Life Cycle
3
Requirements Analysis
 The most important step n implementing a
database system is to find out what is needed —
What type of a database is required for the
business organization, daily volume of the data,
how much data needs to be stored in the master
files etc. In order to collect all this required
information, a database analyst need to spend a lot
of time within the business organization talking to
people, end users and get acquainted with day-today process.
4
Database Design
 In this stage the database designers will
make a decision on the database model that
is perfectly suited for the organization’s
requirements. The database designers will
study the documents prepared by the
analysts in the requirements analysis stage
and then start developing a system that
fulfills the needs.
5
Evaluation and Selection
 Once the data model is designed, tested and
demonstrated, the next phase is to evaluate the
diverse database management systems and choose
the one that is perfectly suited for the
requirements of the organization. In order to
identify best performing database for the
organization, end user should be involved in this
phase.
6
Logical Database Design
 Once the evaluation and selection phase is
completed successfully, the next step n the
database development life cycle is logical database
design. The conceptual design is translated into
internal model in the logical design phase. This
includes the mapping of all objects i.e. tables
design, indexes, views, transactions, access
privileges etc.
7
Physical Database Design
 Physical database design is the procedure of
selecting and characterizing the data storage
and data access of the database. The data
storage depends on the type of devices
supported by the hardware, the data access
methods and the DBMS.
8
Implementation
 In most databases a new database
implementation needs the formation of
special storage related constructs to house
the end user tables. These constructs
typically comprise storage group, table
spaces, data files, tables etc.
9
Data Loading
 Once the database has been created, the
data must be loaded into the database. The
data required to be converting and
migrating to the new database, if the loaded
data is currently stored n a different system
or in a different format.
10
Testing and Performance Tuning
 The next phase is testing and performance tuning,
this phase starts soon the data is loaded into the
database. In this phase, database is tested and finetuned for performance, integrity, access and
security constraints. It is very important that the
database administrators and application
programmers work together during this phase,
because testing and performance tuning happens
in parallel.
11
Operation
 Once the data is loaded into the database and it s
fully tested, the database is than released into
production.
 In operation phase, the database is accessed by the
end users and application programs. This stage
includes adding of new data, modifying existing
data and deletion of obsolete data. The database
administrators perform the administrative tasks
periodically such as performance tuning,
expanding storage space, database backup etc.
12
Maintenance
 Database maintenance phase is very important and it
is one of the ongoing phases in DDLC. Factors such
as new business needs, new information
requirements, acquisition of new data etc will make
it essential to formulate ongoing changes and
improvements to the existing design. The major
tasks in this phase include: database backup and
recovery, performance tuning, design modifications,
access management and audits, usage monitoring,
hardware maintenance, upgradation etc.
13
Growth and change
 Modify the database to reflect changes in the
organization. Monitor database performance
for efficiency and customer satisfaction.
14
Database and
Application Development Life Cycle
Process Modeling
Logical Data Modeling
Database Design
Database Generation
Approaches to Modeling
 Top-down modeling
 Bottom-up modeling
 Targeted modeling
Database
Database
Synchronized
Models
Synchronized
Models
Database
Top-Down Modeling
Top-Down Modeling: Designing a New Database
1.
2.
3.
4.
5.
6.
Business information
One process model (DFD)
One logical data model (ERD)
One multidimensional model
One or more relational models
One or more physical models for
each relational model
Database
Bottom-Up Modeling
Bottom-Up Modeling: Modifying an Existing Database
1. Produce a relational model.
2. Modify the relational model and
create additional relational models.
3. Reverse engineer the logical model
from the relational model.
4. Modify and check the design rules
for the logical model.
5. Generate the modified DDL code.
Database
Targeted Modeling
Targeted Modeling: Maintaining Existing Models by Adapting to
New Requirements
1. Change the logical data model.
2. Engineer to modify the relational
model.
Synchronized Models
2. Engineer to modify the logical data
model.
1. Change the relational data model.
Database
What Is a Process Model?

A process model documents the processes that the business
performs and shows how the data flows through processes
between external sources and information stores.

The diagram for a process model is called a data flow diagram
(DFD).
Why Create a DFD?
Data flow diagrams help to:





Document project boundaries
Understand current area of interest
Find inefficiencies in the current process
Communicate complex ideas and processes with users
Determine data flow identification for data modeling
Validating Your Data Flow Diagram
DFD Rules: Process
 Every process has to have at least one input and one output.
 Use verb phrase labels.
 Each process has a unique name.
DFD Rules: External Agents
 External agents move directly to and
from a process.
 Use a noun phrase label.
DFD Rules: Information Store
 Data cannot move directly
from one information store to
another information store.
 An information store cannot
move directly from or to an
external agent.
 One end of an information
flow must be a process.
 Use a noun phrase label.
DFD Rules: Information Flow
 An information flow has only
one direction of flow between
objects (not a double arrow
line).
 An information flow to an
information store creates,
updates, or deletes data.
 An information flow from an
information store retrieves data.
 Use a noun phrase label.