Key Characteristics of DBMS: Advantages Over Traditional File Systems

hello everyone and welcome back continuing with the previous video

introduction to dbms let us see in this video the characteristics of dbms

as i mentioned in the previous video before dbms we were using the file system approach

in traditional file systems data was stored in files now for example

if i have one user say the accounts department who keeps track of the student fees

and their dues this user just needs the student role number student name the fees paid by the students as well as

any dues and if i have another user say the exam department

who keeps track of the grades received by the students that is it can have student role number

student name course and the grades received by that student in a particular course

here both the users are interested in data about students but they have to maintain different

files because both have different interests the accounts department is only interested in the fee details

whereas the exam department is only interested to know about the grades received by the student in each course

so since both have different interests they cannot have a common file and hence have to maintain different files

as we can see here the student details as the student role number and the student name

is being repeated in both the files this leads to redundancy or duplication of data as well as wastage

of storage space whereas in the dbms approach the entire data is stored in a single repository

and multiple users can access the data based on their interests like for example

data like student role number student name the fees paid fees due the grades they have received etc

can be stored in the single repository and any number of users can access this repository

and view whatever data they want hence there is no duplication of data so that is why we prefer dbms over the

traditional file system approach hope you all have understood the difference between the file system

approach and the dbms approach now we are going to discuss

the main characteristics of dbms approach the first one is self-describing nature

of a database system the second characteristic is insulation between programs and data

and data abstraction the third characteristic of dbms approach is support of multiple views of the data

and the last characteristic that we are going to discuss would be sharing of data and multi-user

transaction processing let us look into each of these in detail

the first one self-describing nature of a database system a database system not only consists of a

database but it also contains the metadata we have already seen what a metadata is

in the previous video so it is nothing but the database definition or a complete description of

the database information like the data type and the constraints etc all these

information is what we call as metadata and this information

is stored in the dbms catalog and this catalog is used by the dbms software

as well as it is used by the database users who wants to know what the structure of the database is

let us just see an example of a catalog so that you understand better

in the previous video we have taken this very simple example of a university database that stores

student and course information so based on this very simple example let us see an example of a catalog

so here we'll see an example of a database catalog where relations or table name along with

the number of columns are mentioned in our example of university database we have three tables or three relations

the first table that we have is the student table with four columns then we have the course table with three

columns and then finally the grade report table with

three columns and also in our catalog we can have another table which defines the columns

present in each table or in each relation their data type

and to which table or to which relation that column belongs to based on our example of university

database the column student name has the data type character of size 30 and it belongs to the

relation or belongs to student table similarly the column roll number is of integer data type and belongs to student

relation or student table so this is an example of a database catalog

where information about the database or a description of the database is stored now going back to the characteristic

that we were discussing now the next point is dbms software must work equally well with any number

of database applications we can have any number of database applications like for

example a university database a library database or a banking database

as long as the database definition or the metadata is stored in the catalog dbms software

has to work efficiently or equally well with any number of database applications whereas in the traditional file

processing data definition or what we call as metadata in dbms

is part of the application programs itself so they work with only one specific

database application because the structure is declared in the program

itself unlike dbms they cannot work with more than one database application so we have seen the

very first characteristic of dbms approach now moving on the next characteristic of

dbms approach is insulation between programs and data and data abstraction in traditional file

processing the structure of data files is embedded in the application programs

itself that is the structure is within the program it is not separate so if there is any

change to the structure of data files it is going to affect the programs we also need to change the application

program whereas in the database approach the structure of data files is stored in the dbms

catalog and it is separate from the access programs since the structure is stored separately

from the programs so if there is any change brought to the structure of the data files it does not

affect the programs and that is called as program data independence we have already seen an

example of university database so in that database we had a student table or a student record

so let us see an internal storage format for that student record here we have the data item name or the

column name in that record or in that table the starting position

and the length in characters in traditional file systems this structure is present within the

programs itself now say if i want to add another piece of data say the date of birth

since the structure is changed this program will not work and the program has to be changed

whereas in the database approach if you want to add a new piece of data say the date of birth

we can add it without affecting the programs since the structure is separate from the

programs we don't have to change the programs we just have to add this new information

in the catalog so next time a dbms program refers to the catalog

the new structure will be referred and used so that is called as a program data

independence now the characteristic that allows program data independence

is called data abstraction dbms provides users with conceptual representation of data

that is it hides the behind the scene details or the complicated details that are not of interest to the database

user a data model is a type of data abstraction

that provides conceptual representation to the database users in a way that most users can

understand and the data model hides the storage and implementation details from the

users so this is what is insulation between programs and data or what we call as program data independence

and the characteristic that allows the program data independence is called as data abstraction the next characteristic

of dbms is support of multiple views of the data a database has many users and

each of those users will have different needs or they may require a different view or

a different perspective of a database like for example

one user will be only interested to know the grades received by the students while another user will be interested to

only know the courses taken by each student so dbms will provide only that necessary data that is

required for each user a view which is a subset of database

that is it is derived from a main database table it contains virtual data or

data that is not physically stored or that is not explicitly stored so that is

support of multiple views of the data where dbms will provide as many views as required by each

database user the next characteristic of dbms is sharing of data

and multi-user transaction processing as the name suggests a multi-user dbms allows multiple users to access the

database at the same time now for multiple users to

access the database at the same time dbms has to provide or it must include concurrency control now what is

concurrency control concurrency control is nothing but when multiple

users are sharing the same database at the same time it should prevent two users from

editing the same data at the same time like for example we have travel agents or

reservation agents and if one of the agents is booking a bus ticket for a particular

passenger and if he is assigning a seat for that passenger

then that particular seat should get blocked and no other agents should be able to access or

book that seat for any other passenger and that is what we call as concurrency control these kind of

applications that require concurrency control are called as oltp or online transaction

processing and it is a major part of database application dpms must enforce or must ensure

several transaction properties and here transaction means program under execution so the two

different transaction properties are isolation and atomicity now what is isolation like we already discussed in

the previous example if one agent assigns a seat to a particular passenger

then that seat is blocked or it is isolated from other agents so that is one of the

transaction properties that dbms must enforce the next property is atomicity where

dbms should ensure that the transaction is executed completely

or none at all now when an agent is assigned a seat for a passenger and if the transaction is stuck midway

or if it is incomplete then that seat has to be released for

other agents to access so either the transaction should be executed completely or none at

all so this is sharing of data and multi-user transaction

processing with this we come to the end of this video hope you all have understood the main

characteristics of dbms approach thank you you