Introduction - Database Processing: Fundamentals, Design, and Implementation - page 17

and other administrative functions. DB2 is a DBMS product from IBM. Most people would
agree that it has faster performance than SQL Server, that it can handle larger databases,
and that it is also more difficult to use. Finally, the fastest and most capable DBMS is Oracle
Database from Oracle Corporation. Oracle Database can be configured to offer very high per-
formance on exceedingly large databases that operate 24/7, year after year. Oracle Database is
also far more difficult to use and administer than Microsoft SQL Server.

Database Design

Database design (as a process) is the creation of the proper structure of database tables, the
proper relationships between tables, appropriate data constraints, and other structural compo-
nents of the database. Correct database design is both important and difficult. Consequently,
the world is full of poorly designed databases. Such databases do not perform well. They may
require application developers to write overly complex and contrived SQL to get wanted data,
they may be difficult to adapt to new and changing requirements, or they fail in some other way.
Because database design is both important and difficult, we will devote most of the first
half of this text to the topic. As shown in Figure 1-18, there are three types of database design:

●

Database design from existing data

●

Database design for new systems development

●

Database redesign of an existing database

Database Design from existing Data
The first type of database design involves databases that are constructed from existing data, as
shown in Figure 1-19. In some cases, a development team is given a set of spreadsheets or a set
of text files with tables of data. The team is required to design a database and import the data
from those spreadsheets and tables into a new database.
Alternatively, databases can be created from extracts of other databases. This alternative
is especially common in business intelligence (BI) systems, which include reporting and data
mining applications. For example, data from an operational database, such as a CRM or ERP
database, may be copied into a new database that will be used only for studies and analysis. As
you will learn in Chapter 13, such databases are used in facilities called data warehouses and
data marts . The data warehouse and data mart databases store data specifically organized for
research and reporting purposes, and these data often are exported to other analytical tools,
such as SAS's Enterprise Miner , IBM's SPSS Data Modeler , or TIBCO's Spotfire.

Figure 1-18
types of Database Design

• From existing data (Chapters 3 and 4)
Analyze spreadsheets and other data tables
Extract data from other databases
Design using normalization principles
• New systems development (Chapters 5 and 6)
Create data model from application requirements
Transform data model into database design
• Database redesign (Chapter 8)

Migrate databases to newer databases

Integrate two or more databases

Reverse engineer and design new databases using

normalization principles and data model transformation

Note: Chapter 7 discusses database implementation using SQL. You need that knowledge
before you can understand database redesign.