Structure and Functions of Computer Database Systems

Attahir Yusuf and Godson Nnadi

Commercial data processing has been concerned

with structured routine business procedure areas
such as stock control, payroll, and invoicing.
These routines entail vast amounts of data essen-
tial for managerial decision making and other
purposes. The data may be in the form of reports,
records or general information. The advent of com-
puter data processing gave rise to the need to store
these data centrally for manipulation and review.
Thus emerged the database.
In the early days, the user manager could in-
The effectiveness and efficiency with teract with the data only through the data proces-
which commercial organizations process sing staff and was not concerned with the prepara-
large quantities of data they generate is tion of data.
becoming a key competitive weapon. With the invasion of business organizations
by microcomputers and the ability of linking them
Computer system development in with the mainframes, managers are using termi-
recent years has led to the creation of nals to interact directly with the database and are
databases—easily updated filing systems getting used to computer-derived materials. This
trend has increased the demand for easy access to
that provide easy and rapid access to a data. The need for a system that provides easy ac-
number of users with diverse needs. cess and also ensures the integrity of the data has
increased. These developments have led to the cre-
Yusuf and Nnadi explain what ation of Database Management Systems (DBMS).
database systems are, their functions,
In what follows we:
trends in their structures, and how they
should be evaluated. • discuss database concepts and components
• its structure and relevance as a centralized
system
Attahir Yusuf and Godson Nnadi are • evaluate a system of managing the database
lecturers at the University of Maidugiri, that will ensure effective maintenance,
Nigeria. independence, and the integrity of the
database
• show how a database can alter managerial
practices.

Database System
Today almost all business concerns, whether small
or big, have some kind of a database. The motor
spare parts dealer when asked for spares consults
his database, the Nigerian Airways Booking Clerk
will consult his database, i.e. the flight schedule, different departments of the organization to get
if you want to book for a flight; a company filing this information. That would be time consuming.
cabinet and the Nitel telephone directory are also But with a computerized database, he can obtain
databases. Any organized collection of informa- the required information within seconds of his re-
tion is a database. Databases are not restricted to quest from the corporate database via a terminal.
computer databases alone. Because of the volume The corporate database today is fundamental
of data in our organizations, the use of the compu- to all commercial data processing. The prolifera-
ter to store data is inevitable. In this paper, we are tion of microcomputers and the advent of
concerned with computer database systems. spreadsheet software (which is capable of answer-
A database is defined as: ing "what if" questions), have given additional
impetus to the database trend. The database has
"A computer filing system that uses a particu- come to occupy a central position in the organiza-
lar filing organization to facilitate rapid up- tion because it is related to almost every system
dating of individual records, simultaneous function and component.
updating of related records, easy access to all
records by all application programmes, and
rapid access to all stored data which must be Decision to Instal
brought together for a particular routine re-
port or inquiry or a special purpose report or The concept of the integrated database may look
inquiry." (Scott, 1986) simple, but it is among the most complex of com-
puting systems. It is necessary for management to
It follows from this definition that a database is a critically and realistically evaluate any attempt to
multi-functional computer system which or- instal a database, for not all companies require a
ganizes files in such a manner that a form of as- computer database. A decision to implement a
sociation exists between different records in the corporate database system cannot be taken in iso-
database. An individual entry of a transaction is lation because of its relationship to various func-
automatically reflected in all the other records in tions of the organization. When the database
the system that are affected by the transaction. It system was invented early in the 1970s (it was in
implies that the database is an integration of many the context of large computers), many corpora-
files. By integration we mean that the database has tions were carried away by the glamour of the new
the ability to make all files act as one. The database technology and installed the system without first
is also continuously updated to answer all unanti- ascertaining whether they really needed it. It is not
cipated queries. surprising that later, some of these corporations
A database, therefore, is a collection of in- were disappointed.
tegrated and inter-related data files, well-orga- The buyer must evaluate not only the need for
nized for multiple usage by multiple users. It is the database system, but also the capacity of the
maintained independently of the user program- system to meet his demands. The buyer must also
mes, thereby providing a rational method of ad- be clear about the result sought. Usually, the ex-
ding, updating, manipulating, and deleting prog- pected result is a central pool of information that
rammes. can constantly be updated and from which users
from different levels and units can draw data to
Corporate Asset meet their needs. This information pool should
also provide all elements for application and still
Over the past few years, companies have come to retain its independence.
realize that a corporate database is one of its most
important assets. This has increased managerial The buyer must also not underestimate the
interest in database operations. Consider the fol- cost involved in installing the system, and must
lowing example. Before accepting an order, a also appreciate that although the database has
salesman would need information on the availabi- functional flexibility, its implementation involves
lity of stock, the production schedules, the availa- substantial design rigidity.
bility of materials and so on. This information will To justify the installation of a database
enable him not only to know whether the product system, a firm should have a large number of on-
is in stock but also if it can be produced. Without a line data retrieval terminals and the ability to
database, the salesman would have to go to the write special programmes for particular informa-

38 Vikalpa
tion needs. It also requires a stable database with a cent innovations in the database system—the rela-
stable format. The stability of the format is neces- tional database—have prevented it from becom-
sary because even the simplest change may require ing one.
re-formating all data files. And this may be very
complex and expensive. When the traditional database system—
hierarchical and network—were developed, the
cost of storing data and processing were prohibi-
Structure and Functions of Database tive. It is understandable, therefore, that they were
Management Systems concerned with efficiency in the storage of data
and conservation of processing power. The
The structure of database system has changed a lot systems were also dedicated to physical location
over the years. of data. Access was by professional data proces-
sing staff through procedural computer languages,
Hierarchical Database requiring a step-by-step process of invoking and
using the system.
When the system was first introduced in the
1970s, it was hierarchical, with a branching tree Relational Database
structure. This system had some functional dif-
ficulties. Before a user could gain access to the The 1980s have brought a new database system
system, he needed to have a detailed idea of the called the Relational Database—the easiest to use
structure since it was interdependent and inter- of all the database systems. It was developed by an
related, so that any action in one sub-system IBM researcher Tedd Codd as an answer to com-
would generate some reactions in the other sub- mercial users of database who are today asking
systems. They were, therefore, used in a batch how they can use and manage the millions of in-
mode under the control of skilled- data processing structions per second that the mainframes of the
personnel. The hierarchical structure is still com- 1990s will provide.
mon today. A diagrammatic view of a hierarchical Relational database is a powerful software
database is depicted in Appendix 1. which organizes data into a table of rows and co-
lumns according to identical elements. The soft-
Network Database ware also has a dictionary organized in a relational
format used as a reference for the content of the
The hierarchical system was followed in the 1970s datapool. The system allows access to the database
by what is known as Network Database or Inter- independent of the storage and indexing techni-
related File System. This database system allowed ques. As a result, it is possible to access all records
interlinking between related items of data; thus in the computer in a random manner.
each record could have more than one parent (BIM,
1986). The network system is efficient in comput- Appendix 3 provides an illustration of the
ing; it can be used on-line. It was an improvement structure of a relational database. For example, a
over the hierarchical system in that one did not person who is five years from retirement and con-
have to wait for a batch to run before accessing tributes to the National Provident Fund (NPF) will
data. Appendix 2 provides a diagrammatic illust- have his name showing up in at least four lists in
ration of the network database. the relational database. The name will appear un-
der factory workers, male or female, NPF con-
A large number of network databases are in tributors, and, those workers who are five years
use all over the world. However, the network from retirement.
database suffers from one major drawback. The
system requires some degree of professionalism in The database is named relational because it is
data processing before it can be used. Profession- based on a predetermined relationship between
als, known as Database Administrators, have to files within a database structure and between
manage data as a resource so that maximum value fields and records in files. They are addressed by a
is gained from the database. Also, effecting fourth generation computer language, which is
changes in the system is a difficult and complex English-like. This means that the end-users can
task because a high degree of system tuning is re- have easy access to information without the
quired to maintain performance. Although the services of data processing personnel. Also, re-
network was developed as a standard database, re- trieval, storage, and manipulation functions are

Vol. 13, No. 4, October-December 1988 39

carried out using names of files or records. Thus, Functions of Databases
the user is not concerned with the mechanics of
the search, as this is done by the system. Databases can be categorized into two based on
their functions—Operational and Informational
From the foregoing, it is apparent that the re- databases.
lational database is a multi-key indexed system
providing multiple relationship between data Operational Databases. Operational databases
(Sharpe, 1987). . were the first to be used and still are the most com-
mon. Their function hinged on batch processing
Under the traditional hierarchical and and handling of data in certain areas of business
network database systems, system designers have operation like payroll, sales, general accounting
to know beforehand the relationships users want and production.
before they can create the database. This is be-
cause it is very expensive to change the format of Operational databases are generally con-
the database. Relational database system was a sidered efficient and work best where the data are
new technology of indexing that provides easy ac- loaded in bulk and also retrieved in bulk form. A
cess. Data can be changed by adding or deleting disciplined mode of data entry is necessary.
even after the data is loaded into the system. How- Operating the system requires devoted data pro-
ever, despite these differences between traditional cessing personnel with the ability to write extra
and relational databases, the distinction between programmes that may be needed for generating
them cannot be said to be clear-cut. Relational special reports and for maintaining system
technology is making in-roads into the traditional efficiency.
system.
Informational Databases. The informational
Many organizations have already installed databases are primarily meant for providing infor-
and invested a lot on either hierarchical or mation on demand to a variety of users. The de-
network database. Rather than scrap them in velopment of the informational database is rooted
favour of the new efficient and simple relational in the realization by management that the database
database system, established software vendors should not only provide historical reports, pro-
have developed interfaces which clone the tradi- duction schedule or payroll, format but all infor-
tional database into relational systems. Commer- mation that may be required for managerial func-
cial users are now increasingly turning to these tions like planning and organizing. It became clear
interfaces as the most viable way of preserving that immediate access to information leads to
their database investment while taking advantage timely and effective decisions. It is in this context
of the latest relational technology. that the database is the input to Decision Support
System.* The informational database should be
One major drawback of the relational database characterized by ease of access, and have the abil-
is its excessive demand for computer processing ity to bring together all loosely connected data.
power and consequent slow response. This means For instance, it should enable bringing together
that the financial investment required by the internal data on productivity and absenteeism and
system is heavy, and many organizations may not external data like dates of football fixtures. (BIM,
be able to instal the system. It is advisable, there- 1986).
fore, that companies should not ignore the tradi-
Requirements for operational procedures have
tional systems. Where they have already installed
remained fairly static over the years, unlike those
a traditional system, companies should explore
for informational databases. The demand for infor-
the possibility of adding an interface to make the
mational databases has been on the increase with
system a relational clone. Alternatively, the com-
panies, especially smaller ones, can fall back on the proliferation of microcomputers. The evolutio-
nary process of information technology is far from
the cheaper microcomputers and use microcompu-
complete. It is anticipated that by the early 1990,
ter software like DBASE III. However, for a large
artificial intelligence will have a great impact on
corporation this would mean decentralizing the
the fourth generation computer languages which
database. For medium-sized companies, micro-
computers may be adequate, especially with a uti- * Decision Support System is a model-based approach to man-
l i t y linking multiple microcomputers into a agerial activity aimed at assisting a manager in decision mak-
network. ing by processing data and making judgement.

40 Vikalpa
will enhance the usability of databases. growing presence of microcomputers, DBMS sup-
From the foregoing, it should be apparent that pliers are urgently looking for other markets.
the relational systems are best suited for informa- There are a variety of DBMS software being offered
tional databases, while the traditional systems fit both for mainframes and microcomputers. Their
the operational databases. However, in practice, prices range from, say, 2,000 Naira (about Rs 2,000
both requirements can be provided in the same at Rupee = Naira) for the microcomputers and up
database. to, say, one million Naira for the mainframes. In-
stallation, testing, and data transfer may cost as
much.
Database Management Systems (DBMS)
DBMS software are offered by both computer
We have presented a database system as a collec- manufacturers, as custom built to match their
tion of inter-related files, logically arranged for ac- hardware, and by independent suppliers—especi-
cess by a number of users. In this context of de- ally for IBM and its clones, and for the DEC VAX
velopment of information and data processing System. For the microcomputer, the DBMS may
technology, it is clear that databases need to im- have restricted capabilities. Because of limited
prove and undergo some changes. Also, the ex- processing power, such DBMS may have reduced
panding base of microcomputers and the develop- file access capabilities and no query language at
ment of user-friendly applications software has in- all. Examples of DBMS for the microcomputers in-
creased the demand for direct access to stored clude DBASE II and III, and DATABASE PLUS by
data. This has increased the need for a system that Ashton Tate. For the mainframes, we have INGES
ensures that changes in the database do not affect from Relational Technology Inc. (RTI), which is of
other programmes, that data files can be merged relational file type; IMS for IBM which is
without the need for re-writing the programmes, Merarchical, and IDMS from Cullinet, which is a
and that edited data is available for all applica- Network System. Others include TOTAL from
tions. Security and maintenance of integrity of Cincom, ADABAS from Software AG, and RAP-
data also become necessary with the increase in PORT from Logica. FOCUS and ORACLE are for
the frequency of direct access to databases. A soft- both the microcomputers and the mainframes.
ware system called DBMS has been developed to IBM has been extremely influential in the de-
take care of these requirements. velopment of DBMS. Software development is
Database Management Systems are generally aimed at the IBM. For instance, the latest software
described as utilities or applications software; from Cincom, the SUPRA, with nearly 250 million
they are programmes required to organize, man- Naira in development investment, and expected to
age, merge, or sort data. A DBMS manages the ac- be the first truly relational system to become com-
cess to, and manipulation of a corporate or an indi- mercially available for the MS-DOS system envi-
vidual's database, assuring at the same time, the ronment is aimed at IBM.
independence of data in the database. In essence,
provision is made for sharing of data among many Selecting Software
applications. Additionally, the DBMS performs
the function of ensuring data integrity and data Because of the cost involved in developing a
security. We shall now consider the DBMS soft- DBMS and problems of compatibility between the
ware available and the steps required to be fol- system and computers, it is necessary that the
lowed while purchasing or installing them, and selection, testing, and implementation of the
explain the intricacies involved in ensuring the system is done by professionals. While choosing a
security and integrity of data in the context of the DBMS, there are factors other than cost that need
role of a DBMS. to be considered. Logically, the system cannot be
selected before defining the objectives for which it
DBMS Software is needed. At the same time, organizational objec-
tives should not be changed to suit a particular
Originally, DBMS was intended for monitoring a DBMS. It should be the other way round. The re-
large volume of data. The implication of this was quirements that need to be evaluated while choos-
that it could be used on large mainframe comput- ing the system can broadly be divided into two
ers only. However, with the number of operating categories—general requirements and special re-
mainframes on the decline, coupled with the fast quirements.

Vol. 13, No. 4, October-December 1988 41

General Requirements. The first thing the user For instance, CROWNTEK 204 runs only on IBM
must consider is the suitability of the system for machines, so a company with a DEC-VAX system
the type of application. This can be done effec- cannot buy CROWNTEK 204, unless it is ready to
tively by identifying the sort of application that is change the hardware to an IBM. This may prove to
involved: is it informational or operational? This be a more expensive alternative.
is relevant because each DBMS is more suited to a
certain type of activity than the others. For inst- Finally, consideration should also be given to
ance, a relational database is more appropriate prospects for ease of installation and modifica-
where large new files are to be created from exist- tions, and potential for networking which would
ing files. The second step is to estimate the volume allow communication between data files in the
of data handled or generated by the organization. database system.
That will determine the level of the DBMS require- From the foregoing discussion on DBMS, we
ments. It will not be economical for a small organi- can see that the effect of its development has been
zation with little source data to purchase the IMS to make stored data more conveniently accessible
which costs nearly 500,000 Naira. The larger the for an increased number of users. It has enlarged
volume, the more sophisticated and expensive the the functionary role of data processing far beyond
system. routine accounting procedures and has provided
For DBMS required as an information system, logical access to stored data and information, and
we need to estimate the needed frequency for up- it is simple enough for even the unsophisticated
dating data. Such systems prefer a low frequency users. It looks like this will be the framework for
of updating, because any updating may cause ma- the "expert systems," that attempt to simulate an
jor changes in the database. The DBMS for infor- experienced man in some limited application do-
mation systems may also have to meet the require- main, and is expected to be one of the features of
ments for ad hoc report preparation and provide fifth generation computers.
for a wide range of previously undefined accesses.
On the other hand, DBMS required for operational Database Security and Integrity
applications should be capable of dealing with Security has always been a problem with
high frequency of data updating. Transactions in databases, but today with the shift from opera-
companies are a continuing phenomenon, and so tional systems to informational databases, security
there is a need for constant updating of the has become a major headache. The complexity is
database. The system should, therefore, be able to three dimensional:
accommodate this requirement. It should also be
able to produce standard reports. • informational database systems tend to
bring in a high level of sharing of the
Special Requirements. This category covers re- database, i.e. an increase in the number of
quirements within the system that will enhance users
security, monitoring, and maneuverability of data. • usage may not be restricted to the comput-
The first is the security requirement. Where there ing personnel. Its on-line nature implies
is direct access to data, especially with informa- that the system can be accessed from re-
tional databases, the issue of security is very im- mote locations, e.g. a manager's office
portant. The DBMS in this case is expected to con- • the addition of new classes of users who
tain a provision for authorized individuals only; are not experienced in computer use.
and a "username" to prevent people from "log-
ging" into other people's files. This is discussed in There is similarly the problem of integrity of
greater detail in the next section. Consideration databases. While security problems like data
must also be given to existing hardware and its sabotage are essentially human problems, in-
limitations. For example, it is not feasible to instal tegrity problems could arise due to human and
a large DBMS in a computer with a small main technical reasons. The danger in the case of in-
memory. tegrity is not theft of data. For instance, hardware
failure may alter the structure or content of the
Systems are different. A particular software database. Other sources of loss of integrity are in-
may not be accepted by a particular hardware. correct data entry by users, inaccurate operation or
Therefore, thought should be given to matching failure of systems software, and improper use of
the system with the hardware that will be used. application programmes (Weldon, 1980). This risk

42 Vikalpa
is particularly tied to the increase in the number of Conclusion
non-professional system users.
We believe that database technology will dramati-
How can these problems be alleviated? The cally influence managerial practices. Database
techniques and processes for ensuring the security systems increase data accessibility thereby trans-
and maintaining the integrity of data are related. forming managerial modes of operation. With a
However, maintaining integrity has a broader database, a manager can react to critical events
scope as it is aimed at guarding the database even quickly with actions based on full information. He
from friendly users while security measures are can also recognize and respond to changing situa-
designed to prevent illegal use of the database tions more effectively. Today's managers should,
only. therefore, address themselves to this new techno-
logy as a means of improving their managerial
To ensure the security of the database, the systems.
DBMS software usually incorporate features that
control access like the password, a means of au- References
thenticating the identity and security rating of the
user. So, if somebody wants to use a database, he HIM (1986). "Data Management Supplement," Management
must have an officially granted password by the Today, March.
appropriate authorities. Before the user can get ac- Dale, C J (1977) An Introduction to Database Systems. Read-
cess, he must provide his password when de- ing, Mass: Addison-Wesley, Second Edition.
manded by the system. The genuineness of the Martin, J (1972). Security, Accuracy, and Privacy in Computer
password will be verified by the system and clear- Systems. Englewood Cliffs, NJ: Prentice-Hall.
ance to use the application programmes and data Scott, G M (1986). Principles of Management Information
files may be granted or withheld. System. New York: McGraw-Hill.
Sharpe, R (1987). "Developing Relationship," Management
The DBMS also restricts access to various Today, December.
levels of information and stipulates what can be Weldon, J (1980). "Maintaining Database Integrity," Working
done at these levels—whether read or write only, Paper Series, CRIS, New York University.
shared or unshared or modify data only.
Usernames are usually given to allow the system
to know what sorts of information and functions
could be provided to the user. The username also
prevents users from breaking into other people's
files. As part of the security procedures, the DBMS
also records type of access granted to users, data
source, etc.
Weldon (1980) has presented two techniques
for the maintenance of data integrity: validation
and consistency. Validation is a procedure of data
and processing control within the system that
checks the accuracy of data. It is designed prima-
rily to prevent the feeding of wrong information to
the databases, by editing all entering data and re-
jecting those found unacceptable. To ensure the
consistency of a database, measures must be emp-
loyed to forestall system failures and introduce
some provision for the recovery of the database
when such failures occur. It is, therefore, neces-
sary that businesses with a database system must
have a back-up which can facilitate the retrieval of
data in the minimum amount of time and with
minimum overheads after a failure. Most DBMS
packages have features that guard, validate, and
recover database as outlined above.

Vol. 13, No, 4, October-December 1988 43

 Appendix 1: A Hierarchical Database

Customers

Appendix 2: Network Datahase

Company Product

Design Budget

Purchase Supply

Sales Advert

Appendix 3: Structure of Relational Database

Name Dept. Age Sex NPF Years of

Service

Pele Finance 30 Male Yes 10

Zico Personnel 45 Male Yes 25
Steffi Finance 22 Female Yes 2
Borg Marketing 30 Male Yes 10