A Detail Database Architecture

For more Https://www.ThesisScientist.com
Unit 2
Database Architecture
Database Architecture :
Database standard proposed by ANSI SPARK in 1975 is being used worldwide and is the only
most popular agreed upon standard for database systems. The Three Level Schema architecture
provides us a number of benefits. For accessing data at different levels we have a number of
users because not all users have to access data in database at all the database levels. The 3 levels
architecture allows us to separate the physical representation of data from the users‟ views of
data.
In the database, same data is stored in a specific feasible format and is available to different users
in different formats as desired by different users. For example, consider we have stored the DOB
(Date of Birth) in the database in a particular format, like in the form of dd-mm-yyyy (for
example, 28-03-1987). However, the users from different departments may require to view the
date of birth in different forms; the examination department may ask it to be displayed as month-
day-yyyy (like march-28-1987) the Registrar‟s office may ask to display date of birth as
mm/dd/yyyy, still the Library may need the in the form of dd/mm/yy. The Three Level Schema
allows us to access the data in different formats at the external level, which is stored in a specific
format at the internal level. The Three levels architecture is useful for hiding the details of
internal systems;

Fig. 1: Three level architecture of database
The Architecture :
The schemas as it has been defined already; is the repository used for storing definitions of the
structures used in database, it can be anything from any entity to the whole organization. For this
purpose the architecture defines different schemas stored at different levels for isolating the
details one level from the other.
Different levels existing pat different levels of the database architecture pare expressed below
with emphasis on the details of all the levels individually. Core of the database architecture is the
internal level of schema which is discussed a bit before getting into the details of each level
individually. The internal level implements all the inner details and defines the intentions of the

database. Internal schema or view defines the structures of the data and other data related
activities in the database. For example it defines that for a student what data will be stored in
terms of attributes of the student and it also defines how different values for these attributes will
be stored, also it tells that who is allowed to make changes to the database and what changes he
can make, etc. These details give us the internal schema and are called the intention of the
database. p(ManMaA7M for pa database is almost permanent, because while designing the
database it is ensured that no information is left behind which is important enough to be stored in
the database and what information is important to be stored in the database from the future point
of view. Once the intention of the database has been defined then it is undesirable to change the
intention for any reason. Because any small change in the intention of the database may need a
lot of changes to be made to the data stored in the database. w,anMiA7M of the database is
performed on the bases of a complete intention, i-e once a database has been defined it is
populated with the data of the organization for which the database is created. This population of
the database is also called as the extension of the database. Extension is always done according
to the rules defined in the internal schema design or the intention of the database.
Effects of changes made to different levels of the database architecture :
We can make changes to the different levels of the database but these changes need very serious
consideration before they are made, Changes at different levels of database architecture need
different levels of users attention for example a change to the data made for the extension of data
will effect only a single record whereas when we make a change to the internal level of the
database the change effects all the stored records, similarly an invalid change in the extension of
the database is not that fatal as a change in the intention of the database because a change in the
extension of the database is not very hard to undo; incase of a mishap whereas a change of the
same magnitude to the intention of the database might cause a large number of database errors
(inconsistencies and data loss).
External View (Level, Schema or Model) :
This level is explicitly an end user level and presents data as desired by the users of the database.
As it is known that the database users are classified on two grounds.

 Section of the organization
 Nature of Job of the users
The external level of the database caters to the needs of all the database users starting from a user
who can view the data only which is of his concern up-to the users who can see all the data in the
database and make all type of actions on that data.
External level of the database might contain a large number of user views, each user view
providing the desired features and fulfilling requirements for the user or user group for which it
is intended. The restriction or liberty a user or user groups get in his rights is the external view of
that user groups and is decided very carefully.
External views are also helpful when we want to display the data which is not lace in the
database or not stored at all. Example of the first case can be a customer Phone number stored in
the database. But when contacting the person it might a wear that the area code for that specific
user is not stored in the database, in that case we can simply pick up the area or city id of the
customer and find the area code for that city from the corresponding Area Codes table. Another
situation may arise when we want to get a student enrolled in an institution and want to make
sure that the student qualifies for the minimum required age limit, we will look the database, for
the students age but if we have stored only the date of birth of the student then the age of the
student needs to be calculated at that very instance; this can be done very easily in the specific
user view and age of the student can be calculated, even the user-view itself can tell use whether
the student qualifies for the admission or not.
As the user view is the only entity or the interface through which a user will operate the database
or use it so it must be designed in such a way that it is easy to use and easy to manage and self
descriptive, also it is easy to navigate through. Also it should not allow the user to get or retrieve
data which is not allowed to the user, so the user view should both be a facilitator and also a
barrier for proper utilization of the database system.
As the system grows it is possible that a user view may change in structure, design and the
access it provides to the users. SO External views are designed
and create in way that they can be modified at a later stage without making any changes in the
logical or internal views.

In the diagram below we can see two different users working as end users having their own
external view; we can see that the same data record is displayed in two entirely different ways.
Fig. 2: Making between External layer and lower layers
Conceptual or Logical View :
This is the level of database architecture which contains the definition of all the data to be stored
in the database and also contains rules and information about that structure and type of that data.
The conceptual view is the complete description of the data stored in the
database. It stores the complete data of the organization that is why it is also known as the
community view of the database. The conceptual view shows all the entities existing in the
organization, attribute or characteristics associated with those entities and the relationships which
exist among the entities of the organization.
We can take the example of the customers of a company. Now the conceptual schema will have
all the details of the products of the company, retailing stores of the company, products resent in
the stock,
products which are ready to be delivered, salespersons of the company, manager of the company
and literally every other thing which is associated with the business of the company in any way.

Now after having all the information we know that the customers buy products from the outlets
of the company, thus in such a case a specific customer has a relationship with that specific
outlet of the company, or the customer may be represented as having association with the sales
person which in-turn has association with the outlet., there may be a number of customers at a
certain outlet and also to mange these salespersons there will be one or more managers. We can
see from the above given scenario that all the entities are logically related to each other in way or
the other. The conceptual schema actually manages all such relationship and maps these
relationships among the member entities. Conceptual schema along-with having all the
information which is to be stored in the database stores the definition of the data to be stored.
The definition may contain types of data, and constraints on data values etc.
Conceptual schema is also responsible for holding the authorization and authentication
information, means that only those people can make use of the database whom we have allowed
to make these changes, so therefore it is the task of the DBMS to ensure be checking the
conceptual schema that he is authorized to check the data or make any changes to the data.
Conceptual schema as it describes the intention of the database; it is not changed often, because
to make a change to the conceptual schema of the database requires lots of consideration and
may involve changes to the other views/levels of the database also.
As in the previous example we saw two database users accessing the database and we saw that
both of them are having totally different user views. Here when we see in the logical view of the
data we can see that the data stored in the database is stored only once and two users get different
data from the same copy of data at the underlying conceptual level.

Fig. 3: External and conceptual layers
By summarizing it all we can say that the external view is the view of database system in which
user get the data as they need and these database users need not to worry about the underlying
details of the data, all these users have to do is to provide correct requirement information to the
DBA or the database designer whoever is designing the database for the system, so that the DBA
or the database designer can create the database in such a way that they can fulfill the users
requirements using the conceptual schema of the database.
Conceptual view/schema is that view of the database which holds pall the information of the
database system and provides basis for creating any type of the required user views and can
accommodate any user fulfilling his/her requirements.
Internal or Physical View/Schema
This is the level of the database which is responsible for the storage of data on the storage media
and places the data in such a format that it is only readable by the DBMS. Although the internal
view and the physical view are so close that they are generally referred to a single layer of the
DBMS but there lays thin line which actually separated the internal view from the physical view.

As we know that data when stored onto a magnetic media is stored in binary format, because this
is the only data format which can be represented electronically, No matter what is the actual
format of data, either text, images, audio or video. This binary storage mechanism is always
implemented by the Operating System of the Computer. DBMS to some extent decides the way
data is to be stored on the disk. This decision of the DBMS is based on the requirements
specified by the DBA when implementing the database. Moreover the DBMS itself adds
information to the data which is to be stored., for this purpose the data compression can be
performed, this space optimization is achieved in such a way that the performance of retrieval
and storage process is not compromised. Another important consideration for the storage of data
at the internal level is that the data should be stored in such a way that it is secure and does not
involve any security risks. For this purpose different data encryption algorithms may be used.
Lines below detail further tidbits of the internal level. The difference between the internal level
and the external level demarcates a boundary between these two layers, now what is that
difference, it in fact is based on the access or responsibility of the DBMS for the representation
of data. At the internal Level the
records are presented in the format that are in match with schema definition of the records,
whereas at the physical level the data is not strictly in record format, rather it is in character
format., means the rules identified by the schema of the record are not
enforced at this level. Once the data has been transported to the physical level it is then managed
by the operating system. Operating system at that level uses its own data storage utilities to place
the data on disk.
Inter Schema Mapping :
The mechanism through which the records or data at one level is related to the changed format of
the same data at another level is known as mapping. When we associate one form of data at the
external level with the same data in another form is know as the external/conceptual mapping of
the data. In the same way when data at the conceptual level is correlated with the same data at
the internal level, this is called the conceptual/Internal mapping.
Now the question arises that how this mapping is performed. Means how is it possible to have
data at one level in date format and at a higher level the same data show us the age. This hidden

mechanism, conversion system or the formula which converts the date of birth of an employee
into age is performed by the mapping function and it is defined in the specific ext/con mapping,
for example, when the data at the conceptual level is presented as the age of the employee is
done by the external schema of that specific user. Now in this scenario the ext/con mapping is
performing the mapping with the internal view and is retrieving the data in desire format of the
user. In the same way the mapping between an internal view and conceptual view is performed.
The figure below gives a clear picture of this mapping process and informs where the mapping
between different levels of the database is performed.
Fig. 1 :Mapping between External/Conceptual and Conceptual/Internal levels
In Figure-1 we can see clearly where the mapping or connectivity is performed between different
levels of the database management system. Figure-1 is showing another very important concept
that the internal layer and the physical layers lie separately the Physical layer 4is 4explicitly
4used 4for 4data 4storage 4on 4disk 4and 4is 4the
responsibility of the Operating system. DBMS has almost no concern with the details of the
physical level other than that it passes on the data along-with necessary instructions required to
the store that data to the operating system.
Figure-2 on the next page shows how data appears on different levels of the database architecture
and also at that of physical level. We can clearly see that the data store on the physical level is in
binary format and is separate from the internal view of data in location and format. Separation of

the physical level from the internal level is of great use in terms of efficiency of storage and data
retrieval.
Fig: 2. Representation of data at different levels of data base Architecture and at the physical
level at bottom
At the internal level we can see that data is prefixed with Block Header and Record header RH,
the Record header is prefixed to every record and the block header is prefixed to a group of
records; because the block size is generally larger than the record size, as a result when an
application is producing data it is not stored record wise on the disk rather block wise which
reduces the number of disk operations and in-turn improves the efficiency of writing process.
Data Independence :
Data Independence is a major feature of the database system and one of the most important
advantages of the Three Level Database Architecture. 4As 4it 4has 4been 4discussed 4already
4that 4the 4file processing system makes the application programs and the data dependent on
each other, I-e if we want to make a change in the data we will have to make or reflect the
corresponding change in the associated applications also.

The Three Level Architecture facilitates us in such a way that data independence is automatically
introduced to the system. In other words we can say the data independence is major most
objective of the Three Level Architecture. If we do not have data independence then whenever
there will be a change made to the internal or physical 4level 4or 4the 4data 4accessing 4strategy
4the 4applications running at the external level will demand to be changed because they will not
be able to properly access the changed internal or physical levels any more. As a result these
applications will stop working and ultimately the whole system may fail to operate.
The Data independence achieved as a result of the three level architecture proves to be very
useful because once we have the data, database and data applications independent of each other
we can easily make changes to any of the components of the system, without 4effecting 4the
4functionality 4and 4operation 4of 4other interrelated components.
Data 4and 4program 4independence 4is 4on 4advantage 4of 4the 43-L architecture the other
major advantage is that ant change in the lower level of the 3-L architecture does not effect the
structure or the functionality on upper levels. I-e we get external/conceptual and
4conceptual/internal 4independence 4by 4the 4three 4levels Architecture.
Data independence can be classified into two type based on the level at which the independence
is obtained.
 Logical Data Independence
 Physical Data IndependenceI
Logical data independence
Logical data independence provides the independence in a way that changes in conceptual model
do not affect the external views. Or simply it can be stated at the Immunity of external level from
changes at conceptual level. Although we have data independence at different levels, but we
should be careful before making a change to anything in database because not all changes are
accepted transparently at different levels. There may be some changes which may cause damage
or inconsistency in the database levels. The changes which can be done transparently may
include the following:
 Adding a file to the database

 Adding a new field in a file
 Changing the type of a specific field
But a change which may look similar to that of the changes stated above could cause problems in
the database; for example: Deleting an attribute from the database structure, This could be
serious because any application which is using this attribute may not be able to run any more. So
having data independence available to us we still get problem after a certain change, it means
that before making a certain change its impact should also be kept in mind and the changes
should be made while remaining in the limits of the data independence.
Physical Data Independence :
Physical data independence is that type of independence that provides us changes transparency
between the conceptual and internal levels. I-e the changes made to internal level shall not affect
the conceptual level. Although the independence exist but as we saw in the previous case the
changes made should belong to a specific domain and should not exceed the liberty offered by
the physical data independence. For example the changes made to the file organization by
implementing indexed or sequential or random access at a later stage, changing the storage
media, or simply implement a different technique for managing file indexes or hashes.
Functions of DBMS
 Data Processing user accessible Catalog
 Transaction Support
 Concurrency Control Services I
 Recovery Services
 Authorization Services
 Support for Data Communication
 Integrity Services

DBMS lies at the heart of the course; it is the most important component of a database system.
To understand the functionality of DBMS it is necessary that we understand the relation of
database and the DBMS and the dissection of the set of functions the DBMS
performs on the data stored in the database.
Two important functions that the DBMS performs are :
 User management
 Data Management
The detailed description of the above two major activities of DBMS is given below;
 Data Processing
By Data management we mean a number of things it may include certain operations on the data
such as: creation of data, Storing of the data in the database, arrangement of the data in the
databases and data-stores, providing access to the data in the database, and placing of the data in
the appropriate storage devices. These action performed on the data can be classified as data
processing.
 User Accessible Catalog
DBMS has another very important task known as access proviso to catalog. Catalog is an object
or a place in the DBMS which stores almost all of the information of the database, including
schema information, user information right of the users, and many more things about the
database. Modern relational DBMS require that the Administrative users of the database should
have access to the catalog of the database.
 Transaction Support
DBMS is responsible for providing transaction support. Transaction is an action that is used to
perform some manipulation on the data stored in the database. DBMS is responsible for
supporting all the required 4operations 4on 4the 4database, 4and 4also manages 4the transaction
execution so that only the authorized and allowed actions are performed.
 Concurrency Support
Concurrency support means to support a number of transactions to be executed simultaneously,
Concurrency of transactions is managed in such a way that if two or more transaction is making

certain processing on the same set of data, in that case the result of all the transactions should be
correct and no information should be lost.
 Recovery Services
Recovery services mean that in case a database gets an inconsistent state to get corrupted due to
any invalid action of someone, the DBMS should be able to recover itself to a consistent state,
ensuring that the data loss during the recovery process of the database remains minimum.
 Authorization Services
The database is intended to be used by a number of users, who will perform a number of actions
on the database and data stored in the database, The DBMS is used to allow or restrict different
database users to interact with the database. It is the responsibility of the database to check
whether a user intending to get access to database is authorized to do so or not. If the user is an
authorized one than what actions can he/she perform on the data?
 Support for Data Communication
The DBMS should also have the support for communication of the data indifferent ways. For
example if the system is working for such an organization which is spread across the country and
it is deployed over a number of offices throughout the country, then the DBMS should be able to
communicate to the central database station. Or if the data regarding a product is to be sent to the
customers worldwide it should have the facility of sending the data of the product in the form of
a report or offer to its valued
customers.
 Integrity Services
Integrity means to maintain something in its truth or originality. The same concept applies to the
integrity in the DBMS environment. Means the DBMS should allow the operation on the
database which are real for the specific organization and it should not allow the false information
or incorrect facts.
DBMS Environments : I
 Single User
 Multi-user
 Teleprocessing
 File Servers

 Client-Server I
 Single User Database Environment
This is the database environment which supports only one user accessing the database at a
specific time. The DBMS might have a number of users but at a certain time only one user can
log into the database system and use it. This type of DBMS systems are also called Desktop
Database systems.
 Multi-User Database Systems
This is the type of DBMS which can support a number of users simultaneously interacting with
the database indifferent ways. A number of environments exist for such DBMS.
 Teleprocessing
This 4type 4of Multi user 4database systems 4processes 4the user requests at a central computer,
all requests are carried to the central computer where the database is residing, transactions are
carried out and the results transported back to the terminals (literally dumb terminals). It has
become obsolete now.
 File Servers
This type of multi-user database environment assumes another approach for sharing of data for
different users. A file server is used to maintain a connection between the users of the database
system. Each client of the network runs its own copy of the DBMS and the database resides on
the file server. Now whenever a user needs data 4from 4the 4file 4server 4it 4makes 4a 4request
4the 4whole 4file containing the required data was sent to the client. At this stage it is important
to see that the user has requested one or two records from the database but the server sends a
complete file, which might contain hundreds of records. Now if the client after making the
desired operation on the desired data wants to write back the data on the database he will have to
send the whole file back to the server, thus causing a lot of network overhead. The Good thing
about this approach is that the server does not have lots of actions to do rather it remains idle for
lots of the time in contrast with that of the teleprocessing systems approach.
 Client-Server
This type of multi-user environment is the best implementation of the network and DBMS
environments. It has a DBMS server machine which runs the DBMS and to this machine are

connected the clients having application programs running for each user. Once a users wants to
perform a certain operation on data in the database it sends its requests to the DBMS through its
machine‟s application software; the request is forwarded to the DBMS server which performs the
required operation on data in the database stored in the dame computer and then passes back the
result to the user intending the result. This environment is best suited for large enterprises where
bulk of data is processed and requests are very much frequent.
 DATABASE DESIGN
Database design and Database Application design are two almost similar concepts, form the
course point of view it is worthwhile to mention that the course is mainly concerned with
designing databases and it concentrates on the activities which are performed during the design
of database and the inner working of the database. The process that will be discussed in this
lecture for development of database is although not a very common one, but it specifies all the
major steps of database development process very clearly. There exist many ways of system and
database development which are not included in the scope of this course. But we will see only
those portions of the other processes which are directly related with the design and development
of database.
Database Application development Process includes the Following Stages or steps :
 Database Design
 Application Programs
 Implementation
These three steps cannot always be considered as three independent steps performed in a
sequence or one after another. Rather, they occur in parallel, which means that from a certain
point onward the application programs development may run in parallel with the database design
stages, specially the last stages of the database design. Similarly while the design phases of the
database are in progress, certain phases of the application programs can also be initiated, for
example, the initial study of the screens‟ format or the reports layout. The database design
process that we are going to discuss in this course does not take these steps independently and

separately, and since the major concern of this course is the design stages of the database, it
concentrate only on those.
 Database Design :
This part of the database application development process is most important process with respect
to the database application development, because the database is something that will hold the
organizations‟ data, in case the design of the database is not correct or is not correctly reflecting
the situations or scenarios of the organization then it will not produce correct result, or even just
produce errors in response to certain queries. So this portion of the database design is given great
attention when designing a database application.
Database Development Process
The database development process means the same thing that we have mentioned as database
application development process. Rather than discussing three stages of database application
development separately, the steps given in the database development process include steps that
cover all three phases mentioned for the database application development process.
Preliminary Study :
Design of database is carried out in a number of steps; these steps play important role in the
design process and need to be given proper attention First Phase of the database development
process is the Preliminary Stage, which is based on the proper study of the system. It means that
all the parts of the systems, or the section of the subject organization for which we intend to
develop the system must be studied. We should find the relation or interaction of different
section of the organization with each other and should understand the way information flows
between different sections of the organization. Moreover it should also be made clear that what
processing is performed at each stage of the system.
 Requirement Analysis :
Once we have investigated the organization for its different sections and the way data flows
between those sections. Detailed study of the system is started to find out the requirements of
each section. This phase is the detailed study of the system and its functionality decisions made

at this stage decide the overall activity of the organization. Requirements of one section of the
organization are fulfilled in such a way that all the sections in the organization are supporting
each other, for example we can say that the results produced by the processing taking place at
one section are used as input for another section. All the users of the systems are interviewed and
observed to pinpoint and precisely define the activities taking place in the different section of the
organization.
Fig : 1. Database Development process
 Database Design :
Third stage in the database development process is the database design; this is a rather technical
phase of the process and need handsome skill as a Database Administrator. This is the phase
where the logical design of the database is created and different schemas for the database are
created logically. Entities are identified and given attributes, relationships are built and different
types of entity mappings are performed.
 Physical Design :
This is the phase where we transform our logical design into a Physical design by implementing
the designed database onto a specific DBMS; the choice of the DBMS is made on the basis of

requirements and the environment in which the system will operate. Implementing a database on
a specific DBMS is very important because it involves the major financial investment of the
organization, and can not be reverted in case a selected DBMS in not capable of providing the
desired efficiency.
 Implementation :
This phase is specific to writing the application programs needed to carry out different activities
according to use requirements. Different users may have different requirements of the data in the
database, so the number of application programs is not known or fixed for all the organizations,
it may vary for different organizations.
 Maintenance of the Database System :
Maintenance means to fine tune the system and check that the designed applications systems are
fulfilling the purpose for which they are meant. Also this phase may involve designing any new
application for the enhancement of the system. Or an already working application may need to be
updated or modified to remove any errors or to add some functionality in the system. The phases
involved in the development of the database application are expressed graphically in Figure-1.
All these stages are necessary and must be given the necessary attention at each level to get
properly working and good system design and a better working environment.
Database Development Process: Approach 2
There are other development processes also with some of the stages or steps modified as
compared to the model we have just studied. Such and alternative is given in the Figure-2 below.
In this design process we see some of the design stages which existed in the previous designing
steps but some of the stages are modified or merged with others to get more precise result or to
distinguish different separate design phases. In this process of designing; the following steps
exist:
 Analyze User Environment
 Develop Conceptual Model
 Map Conceptual Model to Logical
 Choose DBMS
 Develop Physical Design
 Implement System

 Test System
 Operational Maintenance
 Analyze User Environment
This is same step as we discussed while discussing the previous designing process
 Develop Conceptual Model
Next stage in this process model is the development of conceptual model or schema Here we
actually transform the studied and analyzed information into the conceptual design of the
database, this stage may also be connected with the requirement analysis phase, as expressed in
the diagram by showing an arrow from this stage back to the first stage.
 Map Conceptual Model to Logical Model
Third stage is the mapping of the developed conceptual model to the logical model of the
database, means at this stage the schema rules are defined and identified for general database
structures.
 Choose DBMS
Once the mapping of the conceptual and logical model is done, the decision for the use of DBMS
is made; again we refer to the previous model for selecting of the DBMS and will take care of all
the necessary requirements of the environment before making a decision.
 Develop Physical Design
Once we have selected a DBMS, the logical design is then transformed into physical design. This
also includes considering many other decisions, like, data type allocation, indexes to be created,
file organizations, etc. Physical database design is achieved by using the DBMS specific rules for
schema definition and all the facilities provided by the DBMS,

Fig: 3. Database Development Stages. (Second Approach)
 Implement System
This stage is also similar to the one described earlier, i.e., designing the application for
different users and user groups of the organization.
 Test System
Testing is important in the sense that an application may be producing incorrect results,
and this incorrectness may lead to the inconsistency of the system. So when a system design is
complete, once it is implemented it must be tested for proper operation and all the modules must
be checked for their correctness. Whether the system modules are important or not because the
result of the system is mostly dependent on the proper the functionality of all database
applications and modules.
 Operational Maintenance :
Maintenance means to check that all parts of the system are working and once the testing

of the system is completed the periodic maintenance measure are performed on the
system to keep the system in working order.
 Tools Used or Database System Development :
Why tools are used ?
Tools are used for describing the design process in standard ways. If there is no standardized tool
available for designing a specific systems; Then everyone will have to use its own design
notation, and a notation used by one designer may not be understandable to the another one. This
misunderstanding can be more drastic if both the designers are working for the development of
the same system. Tools can also help the designer and the user to mutually agree on a specific
design.
Data Flow Diagrams :
The most common tool used for deigning database systems is Data Flow Diagram. It is used to
design systems graphically and expresses different system detail in different DFD levels.
DFDs show the flow of data between different processes o a specific system.
DFDs are simple, and hide complexities.
DFDs are Descriptive and links between processes describe the information flow.
 Limitation of DFDs
They do not provide us a way of expressing decision points.
DFDs are focused on flow of information only.
 Symbols used in DFD :
There are a limited number of symbols which are used for design process in DFDs.
 DATAFLOW :
The purpose of the dataflow in a DFD is to express the flow of information from one entity to
another entity in the system Data flows are pipelines through which packets of information flow.
Arrows are labeled with name of the data that moves through them Figure-4 below show the
Dataflow diagram

Fig: 4. Dataflow Symbol
 DATA STORE :
Data store is a repository for the storage of the data. When in a system the data is to be
permanently stored somewhere for future reference or use the DATASTORE is used for this
purpose. It is express with a rectangle open on right width and left width of the rectangle drawn
with double lines.
Data in the DATASTORE is held sometimes for processing purposes also i-e it may not be a
permanent data store.. Name of the DATASTORE is a noun which tells the storing location in
the system. Or identifies the entity for which data is stored. Figure-5 shows a data store.
Fig : 5. Data store
 Processes :
Processes are expressed with ovals or rounded rectangles. Processes are used to express the
transformation of incoming dataflow into outgoing dataflow. Process symbols are used for
whatever is the action taking place and whatever is the magnitude or complexity of the action.
Simply stating when data is transformed from one form into another the process symbol is used.
Figure-6a and Figure-6b show two different shapes used for presenting process in DFD.

Fig. : 6a Process Fig-6b
 DFD-Process :
In DFD processes are numbered for expressing their existence at a certain level in the
system.
1.0
Processor
1.0
Processor
Fig: 7. Numbered DFD Processes
 External Entities :
These are the entities interacting with the system in any of two different ways. They may be
either receiving the data from the system, or may be producing the data for the system to
consume.
Shape used to express external entities is rectangle. The shape for external entity is shown in
Figure-8.
Fig: 8. External Entity
 Collector :
This DFD shape is used to express several dataflow connections terminating at a single location.
Collector is used to show the convergence of data to a single point. Fig 9a shows the Collector
symbol and Fig 9b show a collector symbol acting as a sink for multiple data flows.

Fig: 9a Collector Fig 9b. Collector with Multiple Dataflow
 Separator :
The dataflow symbol which is used for separating data from a single source to multiple
sinks is known as a separator.
Figure 10a show the presentation of separator and the figure 10b shows the separator as it
may appear in a DFD.
Fig: 10a Separator Fig 10b. Separator with Multiple Dataflow
 Ring Sum Operator :
This operator is used when data from a source process can flow to one of the mentioned sinks.
For this purpose the symbol used is displayed in Figure: 11a and its presentation in a DFD is
expressed in Figure-11b.
Fig: 11a Ring sum operator Fig 11b. Separator with Ring sum operator
 AND Operator :
This operator is used when data from a source process must flow to all the connected sinks. For
this purpose the symbol used is displayed in Figure: 12a and its presentation in a DFD is
expressed in Figure-12b.

Fig: 12a AND operator Fig 12b. Separator with AND operator
Types of DFD
 Context diagram
 Level 0 diagram
 Detailed diagram
 Context Diagram :
This is the level of DFD which provides the least amount of details about the working of the
system. Context DFDs have the following properties :
They always consist of single process and describe the single system. The only process displayed
in the CDFDs is the process/system being analyzed. Name of the CDFDs is generally a Noun
Phrase.
Fig: 13a. Example Context DFD Diagram

No System details are shown in the Contexts DFDs just context is shown. Input and output from
and to the process are shown and interactions are shown only with the external entities. An
example DFD at context level is shown in Figure: 13a and 13b.
In the context level DFDs no data stores are created. Ant dataflow from external entities are only
directed toward the purported system and vice versa, no communication is show between
external entities themselves.
Fig: 13b. Example Context DFD Diagram
 Level 0 Data Flow Diagrams :
The level 0 Diagram in the DFD is used to describe the working of the whole system. once a
context DFD has been created the level zero diagram or level „not‟ diagram is created. The level
zero diagram contains all the apparent details of the system. It shows he interaction between a
numbers of processes and may include a large number of external entities. At this level it is the
duty of the designer to keep a balance in describing the system using the level 0 diagram.
Balance means that he should give proper depth to the level 0 diagram processes. Because
placing too much details and showing all of the miniature processes in the level 0 diagrams
makes it too much complex. On the other hand it is also not recommended to just ignore

even larger processes of the system, because in such a case although the level 0 DFD will
become simple but now we will have to create large number of detail DFDs. So a balance in
describing the system should be kept so that the depth of the Level 0 DFD is manageable.
Detailed Data Flow Diagram :
This Type of the Data flow diagrams is used when we have to further explain the functionality of
the processes that we showed briefly in the Level 0 Diagram. It means that generally detailed
DFDS are expressed as the successive details of those processes for which we do not or could
not provide enough details.
The symbols and other rules regarding the detailed DFD are same as are in other types of DFDs.
The special features associated with this diagram are that, one, it is optional, that is, it is created
for only those processes from the level 0 diagram for which we want to show the details. For a
small sized system we may not need to develop even a single detailed DFD, since the level 0
diagram might be covering it sufficiently. Second specific characteristic of the detailed DFD is
its processes‟ numbering. Numbering of processes in the detailed DFD is done on the basis of
numbering of the particular process in level 0 diagrams whose sub-processes are being included
in the detailed DFD. For example, a specific process which was numbered in the level 0 diagram
as 1.0 or 1 may have a number of sub-processes since we did not represent the process 1.0 in
detail in level 0 diagrams. So in the detailed dataflow diagram we create sub-processes of that
process and then number all the sub processes of that specific process as the sublets of the
process.
Numbering of such sub processes is done as 1.1, 1.2, and 1.3… for first second and third sub-
processes of the process 1.0 respectively. The phenomenon of creating sub-processes does not
end at creating a few sub-processes for a specific process shown at level 0 diagrams. Rather it
may continue deeper if there is requirement for further explanation of the any process or sub-
processes. In such a case when we create sub-process of a sub-process 1.2 then the numbering is
done in further extension of that specific sub processes number and example of such a
numbering process is 1.2.1, 1.2.2, 1.2.3,…

Another point that is worth mentioning here is that we call processes in the detailed FDs as sub-
processes, but they are sub-processes only in reference to the process whose details they are
explaining otherwise they are just like processes; transforming some input data into some form
of output. The sub-processes may be performing relatively small amount of operations, still they
are processes.
Maximum Number of Process in a DFD should not be very huge. Having a moderate number for
a detailed DFD is also recommended because it adds clarity to our detailed data flow diagram.
For clarity propose it is good to have a maximum of 7 or 9 processes in one detailed DFD.
Moreover all the processes, sub processes, data stores, entities data flows and all other
components of the DFD must be named properly, so that anyone who is using this DFD should
be able to understand the DFD easily.
In all the levels of DFD it must be considered that all the processes have data inputs as well as
data outputs. Data being sent to one process should be processed so that it changes its form and
transforms from one form to another.
When creating a detailed diagram the data inputs and data outputs must be in coincidence, mean
in both the diagrams the data input to a process and data output in the form of data flows must be
same.
Data Dictionary
A database that containing data about all the databases in the database system. Data dictionaries
store all the various schema and file specifications and their locations. They also contain
information about which programs use which data and which users are interested in which
reports.
Types of Data Dictionaries :
 Integrated
There are basically two types of data dictionaries which are available for use by a DBMS,
with respect to their existence.
The first type of data dictionary in this context is the integrated data dictionary. Such a data
dictionary is place embedded into the database system, and is created by the DBMS for its usage
under the directions and requirements provided by the DBA

As the DBMS needs to talk with the “three level architecture” of database and mapping
information along with all the database design information lies in the database schema. The
DBMS uses the data dictionary to access the database at each layer or model, for this purpose the
data dictionary of any type can be used but the integrated data dictionary is far more efficient
than any free standing data dictionary because an integrated data dictionary is created by the
DBMS itself and uses the same data accessing techniques etc.
 Free Standing
Second type of data dictionary is free standing data dictionary create by any CASE tool and then
attached to the database management systems. A number of case tools are available for this
purpose and help user designing the database and the database applications as well in some
modern forms of the CASE tools.
Cross Reerence Matrix
This is a tool available in the data dictionary and helps us in finding entities of the database and
their associations. CRM is developed at the designing stage of the database; we can say that at
the time of creation of the user views of reports for certain users we identify the material
required by the users. In the cross reference matrix, on the Y axis we specify the accessible
components of the database such as transitions, reports, or database objects and on the x axis we
specify the attributes that will be accessed in the corresponding accessed object.
Now the matrix gets a shape of two dimensional arrays on which we have accessible objects of
the database and on the other hand we have the elements which are available for access through
those objects. Then whichever data item is accessible through a certain object we place a tick on
the intersection of that row and column and thus we can easily identify the deferent items
accessed in different reports.
A C c
T S t l C
r e t a a
a m e s s
n R n s s
s e d R R
c s S e e
r C h s s

i a e S u
r e u l
t D t b t
Course Name √ √ √
Cumulative GPA √ √ √
Date √ √ √ √
Father Name √ √
Final Marks √
Grade √ √
GRD Point √ √ √ √
Marks √ √
Mid Term √
Rogram Name √ √ √ √ √
Semester GPA √ √ √
Semester Name √ √ √
Session √ √
Session Marks √
Student Name √ √
Student Name √ √
Student Registration √ √
Table 1: An example cross reference matrix
The cross reference matrix shown in table 1 lists different attributes against different reports
required by different user groups of an exam system. Rows in this matrix contain different
attributes and the columns contain different reports. Now the tick mark in the cells represents the
use or presence of attributes in different reports. This matrix represents, on one side, the relative
importance or use of different attributes. On the other hand it also helps to identify different
entity types and their defining attributes. The attributes that are represented collectively on one
or more reports are candidates of combining into a single entity type. Although it is necessary
that attributes appearing together should be grouped into same entity type, but still they are
candidates for combining into the one.

Data Dictionary in not very necessary for using such a cross reference matrix, instead for
relatively small systems it can be created manually.
Outcome of the Analysis Phase
In the preliminary study phase, database designers collect information about the existing system
from the users of the system. For this purpose they may interview different users or concerned
persons, or they may distribute questionnaires among different users and ask them to fill them in
and later may use these questionnaires in the analysis phase. Designers represent their
understanding of the working of existing system in the form of DFDs and discuss it with the
users to make it sure that they have understood all details of the existing system and the
requirements of different users groups.
The DFDs are input to the analysis phase, where designers analyze the requirements of the users
and establish the procedure to meet those requirements. From the database perspective, in the
analysis phase designers have to identify the facts or data that is required to be stored in order to
fulfill the users‟ requirements. For this purpose they may use some CASE tools, like cross
reference matrix. Generally, in the analysis phase, designers prepare a draft or initial database
design that they ultimately finalize in the next phase, that is, the database design phase. So in
short we can say, that DFDs are the output of the preliminary phase and are input to the analysis
phase. The initial design or a draft form of design (generally in entity-relationship data model) is
the output of the analysis phase and input to the design phase. In the design phase, then you
finalize the design.
Database Design Phase
Database design phase follows the analysis phase. Before starting the discussion on the design
activity, it will be wise if we clearly understand some basic concepts that are frequently used in
this phase.
 Database Design /Database Model
These terms can be used interchangeably for the logical structure of the database. The database
design/model stores the structure of the data and the links/relationships between data that should
be stored to meet the users‟ requirements. Database design is stored in the database schema,
which is in turn stored in the data dictionary.
 Database Modeling

The process of creating the logical structure of the database is called database modeling. It is a
very important process because the designing of the application provides us the basis for running
our database system. If the database is not designed properly the implementation of the system
can not be done properly. Generally the design of the database is represented graphically because
it provides an ease in design and adds flexibility for the understanding of the system easily.
Data Model
Data model is a set or collection of construct used for creating a database and producing designs
for the databases. There are a few components of a data model :
 Structure :
What structures can be used to store the data is identified by the structures provided by the data
model structures.
 Manipulation Language
For using a certain model certain data manipulations are performed using a specific language.
This specific language is called data manipulation language.
 Integrity Constraints
These are the rules which ensure the correctness of data in the database and maintain the
database in usable state so that correct information is portrayed in designing the database.
Generally these components are not explicitly defined in data models, they may be available in
some of the modern DBMSs but in traditional and general model, these may not be available.
Significance of the Data Model
Data model is very important tool because it is something which is sued for designing the
database for a DBMS and no DBMS can exist independent of any data model, now if we use a
specific DBMS but are not sure about the data model it uses for data abase usage, we can not
create a proper database.
As a specific DBMS is base on the use of a specific data model so when using a DBMS it is of
great use to know that what structures, manipulation languages and integrity constraints are
implemented by a specific DBMS. As it is the only way to know the facilities and functionalities
offered by the DBMS.

This is the reason whenever we get a specific DBMS, it is explicitly mentioned with that DBMS,
that which data model this DBMS uses.
Types of Data Models
 Semantic Data Model
These are the data models which provide us better flexibility in implementing constraints, better
language utilities and better data structure constructs. As a result actions performed using proper
data and structure tools gives us better data designing and manipulation facilities. A better data
model provides better opportunities to express multiple situations in the database design and as a
result get better output from the tool or model in the form of a better database design.
 ER- Data Model
 Object oriented data model
 Record Based Data Model
This is the second type of data models available to use and has three basic types
 Hierarchical Data Model
 Network Data model
 Relational Data model
These models are records based and are not in similarity with those of semantic data models.
These models handle the data at almost all the three level of the three layers of the database
architecture. Semantic data models are generally used for designing the logical or conceptual
model of the database system, once very common example of the semantic data model is ER-
Data Model and is very much popular for designing databases. No DBMS is based on ER Data
model because it is purely used for designing whereas a number of DBMS are available based on
OO data model, network data model, relational data model l and hierarchical data model.
Types o Database Design
Conceptual database design
This design is implemented using a semantic data model, for example for creating a design for an
organization database we can use and we do use the ER-Data model.
Logical Database design

This design is performed using a data model for which we have a DBMS available and we are
planning to run our database system that DBMS.
Physical Database Design :
The Logical design created using a specific data model and created after the analysis of the
organization, it needs to be implemented in a physical DBMS software so the physical database
design is performed and the design created so far in the logical form are implemented on that
very DBMS.
By separating the three design levels we get the benefit of abstraction on one hand whereas on
the other hand we can create our logical and conceptual designs using better design tools, which
would have not been possible if we are using the same design-tool for al the three levels.
Moreover if in future there is a need to make a change in the physical implementation of the data
we will have to make no changes in the logical or conceptual level of the database design , rather
the change can be achieved by only using the existing conceptual model and implementing it
again on Physical model using a separate DBMS.

A Detail Database Architecture

More Related Content

What's hot (20)

Similar to A Detail Database Architecture (20)

More from Prof Ansari (20)

Recently uploaded (20)

A Detail Database Architecture