00:06

in this presentation we will see about the database system architecture

00:48
in more data volume than what is specified on the slide so when a
database needs to handle this much of information obviously the main memory cannot
handle this though the main memory is faster in nature but this main memory cannot
store the database because it's volatile in nature at the same time the secondary
storage devices like the hard disk or the ssds the solid state drives will also not
be sufficient due to the increased volume of data then what's the solution for this
because this contemporary world is handling an

01:21
enormous amount of data so obviously we are in the need of storage
servers or data centers or cloud that can handle large volume of data the reason
why i am projecting this is that i'll give you a real time example just see
facebook how much volume of data it is handling every day take another server is a
whatsapp server which is handling an enormous amount of data every day every minute
every second it handles a large volume of data at the same time whatever we post it
is stored in their servers let's say you have

01:52
posted your photo in your facebook account and later you have decided
that you don't want that photo to be posted so obviously we will be deleting that
isn't it so when you delete it's just deleted from the view of yours and your
friends is it actually deleted from the facebook server no it will not so whatever
we upload everything is stored in the servers and think about the complexity of
handling such high volume of data but still our technology and servers are able to
handle this that's the part of being in

02:27
this contemporary world and coming to the dbms architecture this is also
referred as the database system structure and what we are going to learn in this
presentation is the part one of the database system architecture at first let me
show you the architecture diagram here is the database system architecture if you
see the diagram we can see the users are at the top level then iv users the
application programmers the sophisticated users and the database administrators
obviously each users will be preferring some set of tools for

03:00
their activities and here is the tool level and if you see the
architecture diagram we have two important components the query processor part and
the storage manager part we will come to this later and also if you see the
databases are actually stored in the disk so in this example we are assuming that
the database is actually stored in the disk and that's why this storage is referred
as the disk storage so in the database architecture part 1 and part 2 presentations
we are going to exclusively focus on all the individual
 03:32
components that constitute this database system architecture i mean the
components in the query processor as well as the storage manager in part one of
database system architecture we are going to exclusively focus on the storage
manager and the disk storage i mean we are going to focus on the buffer manager the
file manager the authorization and integrity manager and the transaction manager
also we are going to see about the data indices the statistical data and the data
dictionary so let's see one by one in detail so we

04:05
are clear that we are going to focus on the storage manager which is this
so we are in this part and we are going to focus on all these managers before
seeing the managers one by one let's understand why do we need the storage manager
and what role it plays in the database system architecture basically this storage
manager provides the interaction between the low level data stored and the
application programs if you see the architecture diagram this storage manager
provides the interaction between the application

04:36
programs that is from the top level and the data stored at the low level
so here is the actual data that is stored and to be precise here is the actual data
that is stored in the database actually this storage manager only handles all the
queries that are submitted to the system what do we mean by this if you see the
architecture diagram the query processor processes the queries and this will give
the request to the next level which is the storage manager and that is why the
third point says that this is also

05:08
having an interaction that is received from the query processor to the
desk storage and also this storage manager is having the interaction with the file
manager you may be asking me a question we have already compared file system and
database system and we understood that databases have more advantages over file
system in effective storing and retrieval of data even though we are saying that it
is a database system internally it is made up of files only but it's not going to
work as the regular file system rather it is

05:40
going to work differently why because we need to effectively store and
retrieve the data in database management system at the same time the various
drawbacks that we have seen with the file systems needs a solution here though we
are using internally all files but this is not going to work as per the file system
concept it's going to work as per the database system concept and also we know that
only low level data are stored and this low level data is actually the raw data
that are stored on the disk using

06:11
the file system provided by the operating system and obviously whatever
the interaction that we make see we may do some insertion or deletion or updation
whatever it is coming from the front end or from the high level those data have to
be stored in the database to be precise it has to be stored on the disk only
because this database is on the disk only so all the raw data are stored on the
disk and what format it is going to use this is going to use the same file system
format that is provided

06:42
by the host operating system so we are clear that if data are to be
stored on the disk it uses the file system provided by the operating system and
these raw data are provided from the users or from the application layer or from
the top post layer and what about the dml commands what do we mean by dml commands
these dml statements are the statements that deals with the insertion of data or
deletion of data or updation of data or sometimes selection of data from the
databases we know databases are

07:14
going to store the data and data related operations are built by the dml
statements example select insert delete update nowadays in the last lecture of this
chapter we are going to focus on the dbms languages at that time you will be able
to understand what is ddl what is dml and other languages also we have a separate
chapter called sql where we are going to exclusively focus on the sql the
structured query language along with ddl dml tcl and dcl operations i know things
will be unclear but for

07:50
now just understand dml is data related so when such commands are
received this storage manager only translates the various dml statements into low-
level file system commands why because the internal organization needs to be
supported by the operating system and this operating system obviously will be using
some file system and whatever the data manipulation language commands that we give
these things have to be translated into the low level file system commands in
simple terms storage manager is responsible for storing

08:25
retrieving and updating the data now we will see the various managers
that are there in the storage manager as per the architecture diagram we are sure
that these are all the four components or managers in the storage manager part
number one the authorization and integrity manager number two the transaction
manager number three the file manager and number four the buffer manager and also
as already stated we are going to look at various data structures the data files
the data dictionary and the indices

08:57
we will see one by one now but let's go to the architecture diagram at
first we will focus on the first storage manager component the buffer manager and
this buffer manager is actually responsible for fetching the data from the disk
storage here is the disk storage and into the main memory and also this buffer
manager decides what data to cache in the main memory i hope you are already aware
of cache memory in computer systems this memory is mainly used for faster access so
buffer manager takes cares of this at

09:31
the same time this buffer manager is very crucial in terms of another
important operation that is handling the size of the data we know the databases
will be larger than our actual main memory so this buffer manager will also take
care of this memory management and coming to the second one the file manager we
know already this is mainly for allocating the space on the disk storage we know
actually the databases are already stored on the disk and we may be inserting new
data or deleting the existing data or updating the

10:03
existing data this file manager is responsible for allocating the space
on the disk storage and also the data structures we have a lot of data structures
followed in the disk storage and this file manager is used to represent the
information stored on the disk and coming to the next one the authorization and
integrity manager this component or this manager is used to test for the
satisfaction of all the integrity constraints that we enforce on the database say
for example balance should not be less than zero in such

10:35
case if any operation that is carried out on the database where the
balance is going to be zero or lesser than zero that operation needs to be aborted
that should not be permitted because we have given an integrity constraint that
balance should not be less than zero so all these integrity constraints are handled
by this authorization and integrity manager at the same time authorization of data
i mean who can access what that is also actually handled by this authorization and
integrity manager and coming to the last

11:05
manager the transaction manager and this is the very important part
because this ensures that the database remains consistent despite system failures
or any hardware or software failures at the same time this transaction manager
ensures that the concurrent operations carried out on this database are actually
not conflicting with each other i mean concurrent operations are permitted multiple
users can access the shared data item at the same time without conflicting each
other so this is ensured by the transaction manager

11:40
anyway in the next lecture which is the part 2 of the database system
architecture i will elaborate more on this transaction manager we are done with the
storage managers what are all the components that we have seen we have seen about
the authorization and integrity manager the transaction manager file manager and
buffer manager before we sign out let's see the various data structures used what
are all the data structures we are going to focus on the data files actually this
stores the database itself here is the data files

12:11
so here is the actual database that is residing and coming to the next
one which is the data dictionary we have already seen about metadata that is data
about other data that is stored so these metadata information are stored by the
data dictionary and coming to the third data structure which is indices we know
obviously data are stored in the databases and we need to ensure that the data are
accessed fast storage and retrieval are efficient and effective and we are also
sure that databases are handling large volume of

12:44
data in such cases referring to the exact data item that we want say our
data may be in the first location or it may be in the last location but it should
not be the case that databases takes a lot of time to process our request it should
be as fast as it can so in such case indices helps us to retrieve the data faster
anyway in the chapter indexing and hashing i will be talking about indices
elaborately for now just have this this is used to provide fast access to our data
item and if you want an analogy this is like

13:18
index pages in our textbook why do we need index pages in our textbook
suppose we want to refer something we need not go sequentially to refer that if you
go to the index file this index file will take us to the exact location what we are
looking for and the similar concept is applied on the databases for faster access
so what we have seen now we have seen the various data structures like the data
files the data dictionary and the indices one thing we have left what's that the
statistical data so this part is mainly storing the

13:51
statistics about the data so this will also help us to take some
decisions for managing the databases so what we have learned today in part one of
database system architecture we focused on the storage manager and the disk storage
and already we have seen about the various types of users in one of our previous
presentations so in the next presentation we will focus on the remaining part which
is the query processor so i hope the session is informative and thank you for
watching [Music] [Applause]

14:24
[Music] you