| TMA
July 2002
Question No. 1: Discuss the relative
merits and difficulties of applying
data flow oriented design in the
area of design of Database Management
Systems.
Data flow design is just one
of the many types of design methods,
which are possible, to make a
proper plan for the method of
implementing a software project.
The other design methods include
user-interface design in which
how the user interacts with the
software is designed, the component-level
design in which how the inputs
are converted into outputs is
designed, the architecture design
in which the overall structure
of the software is designed.
Data flow design involves understanding
of how date moved from one module
to another in a project. We find
the various modules, and for each
of the modules, we find out the
inputs provided and the outputs
which it provides. We then associated
there inputs and outputs of each
module, with other modules. For
example, module 1 may accept some
form of input from the user and
result in some output. The output
from module 1 is provided as input
to module 2, which does some process
and results in its own output.
This output may be passed as input
to the next module. This process
continues, until we get the final
results.
A database management system
software is generally a collection
of date and a set of program which
help is accessing those date.
The general functions of the DBMS
software include, the ability
to store new date, the ability
to modify the existing date, the
ability to remove the existing
date and the ability to retrieve
the date in the requested form.
There are a large number of DBMS
software through out the world.
Every company now-a-days is storing
its data in one or the other type
of DBMS. They have become almost
compulsory for top and middle
level management people of any
company to take business policy
decisions.
These DBMS software should have
the capability of managing large
amounts of date, with out providing
any sort of problems to users.
To make this feature possible,
in the recent times companies
have started using data mining,
date warehousing concepts. These
latest things make it possible
for proper and efficient maintenance
of large amounts of data of a
company, without any errors or
problems of any sort.
To prepare a DBMS software itself,
the method of storing date internally
become more crucial then any other
task in the development process.
In analysis of such a project,
we find what type of data is going
to be stored and also the volume
of data, which is probably going
to be stored. Basing on this information,
designers will design efficient
data structures to hold the data
of the company. The selection
of data structures is done by
taking into account, the type
of accessing done by the company
personnel
A data flow oriented design methods
will be very helpful, in understanding
the working of a DBMS type of
software, because it allows to
show, how the date moves between
various components. The DBMS contains
various components like the repository,
tables, database drivers, query
processor etc. How the date moves
between these components can be
understood very clearly with the
help of a data flow oriented design
method.
But the point is that, it is
not important where the date is
moving within a DBMS software,
but in which form the data is
moving. This possible by understanding
the data structures which are
used internally to store the entire
data. The date structure being
used can be a simple linked list,
if the data is accessed by the
company personnel only in a sequential
manner, The data structure can
be a binary tree or a AVL tree,
if the data is accessed randomly
and large amounts of data is stored.
So, in the data flow oriented
design method, if we incorporate
the details if the data structure,
then it will be very useful to
understand the underlying structure
of any DBMS software.
Cont...
|
