Database Reference
In-Depth Information
TABLES
COLUMNS
ROWS
PHYSICAL
DESIGN
LOGICAL
DESIGN
PRIMARY
KEYS
FOREIGN
KEYS
RULES &
CONSTRAINTS
Figure 11-3
Output from logical design as input to physical design.
•
In the logical design phase, you can focus on designing the tables, segments, or
nodes without being bogged down by hardware considerations.
Input to Physical Design
When you finish the logical design phase, you produce a representation of the infor-
mation requirements in the form of tables, segments, or nodes. If you are designing
for the relational data model, your output from the logical design phase consists of
two-dimensional tables with their columns and rows. The logical design phase takes
the database development effort to this stage.
However, you know that data do not exist as tables, or segments, or nodes in
physical storage. Therefore, to implement the target database system, you have to
define the representation of data in formats that are suitable for physical storage.
How do you accomplish the design of the physical structure? The output of the
logical design phase serves as the input to the physical design phase. In the physi-
cal design phase, you take each table from the output of the logical design phase
and transform it into file and record layouts.
Figure 11-3 illustrates how the output from the logical design phase is used as
input to the physical design phase. Note the components of the logical design
output. Also, observe once again how design is simplified by separating it into logical
and physical design phases.
Ensuring Design Completeness
When you are near the end of the logical design phase, usually it is a good place in
DDLC to ensure that the design is complete. You started from the real-world infor-
mation that needs to be represented in the target database system. You gathered
requirements and defined the content, extent, and scope of the information require-
ments. On the basis of the requirements definition, you created a semantic data
model to represent the real-world information. Then you transformed the seman-
tic data model into a relational data model. If your target database system were
hierarchical or network, you would have transformed the semantic data model into
a hierarchical or network data model. If you had adopted the traditional method of
Search WWH ::
Custom Search