OPERATING SYSTEM

OPERATING SYSTEM
Written By: ATUL KUMAR

Written By: ATUL KUMAR Page 1

 OPERATING SYSTEM

➢ Operating system (OS): An operating system is a kind of

system software which consists of a set of programs. It
manages all the computer hardware and provides a basis
for application programs and acts as an interface between
user and computer hardware.
An OS provides an operating environment in which a user
can easily interface with the computer to execute a
program. The OS controls and co-ordinates hardware and
various application programs for various users.

User-1 User-2 User-3 User-4 User-5

System Software and Application Software

Operating System
Computer Hardware

Fig: Abstract view of component of computer

An operating system loads itself into memory and the

loading of OS in memory is called “Booting”. The booting
process starts the moment the computer is switched on.
A firmware called bootstrap program stored in ROM is
responsible for loading the OS into memory. The moment
the computer is switched on, the software on memory
chip is executed to run the device is called firmware.
There are two ways to interact with OS.
a) By using direct command
b) By using system call

Written By: ATUL KUMAR Page 2

 OPERATING SYSTEM

An OS provides commands that let the user communicate

with the OS. For example, Disk Operating System (DOS)
provides commands like DIR, CLS, RD, MD, CD etc. that are
directly executed by user. Similarly, UNIX/LUNIX OS
provides commands like clear, cd, mkdir, cp etc. to
command directly with OS.

➢ System call: It is a set of functions defined in low level

language which act as interface in between application
program and OS.
A computer program requests the service from the
kernelof OS through system call.
In other words, system call provides services to application
program. It provides interfaces between a process and OS.
All programs requiring resources must use system cell.
There are 5 categories of system cells.
I. Process control
II. File management
III. Device management
IV. Information management
V. Communication
❖ Classification of operating system:
An OS which broadly classified into 2-categories:
a). Single user OS
b). Multiuser OS

Written By: ATUL KUMAR Page 3

 OPERATING SYSTEM

➢ Single user OS: - An OS which can handle the request of

only one user and can manage one set of I/O devices,
processor, memory etc. is called a single user OS. Single
user OS performs the task of only one user at a time and it
is specially used to run a stand-alone computer.

• Subdivision of single user OS:

a). Single user Single tasking
b). Single user multitasking

• Single user single tasking: A single user OS which does

only one task of a user is called single user single tasking
OS. This OS can load only one program to memory at a
time. Ex: MS-DOS (Microsoft disk operating system).

• Single user multitasking: A single user OS which performs

multiple tasks at a time of a single user is called single user
multitasking OS.
This OS can load multiple programs intomemory at a time.
When one program waits for some I/O then the OS starts
executing another program loaded in memory. This way a
user gets the illusion that the multiple tasks being done
simultaneously. All GUI OS are single user multitasking OS.
Ex: MS-Windows 95/98/XP/vista/07/08/10, MS-Windows
NT Workstation, MS-Windows 2003 professional etc.

Written By: ATUL KUMAR Page 4

 OPERATING SYSTEM

➢ Multiuser OS: An OS having the ability to handle the

request of multiple users is called multiuser operating
system.
A multiuser OS can support more than 1-set of I/O devices.
, processors memory etc.
A multiuser OS is also called Network OS because it is used
in a network environment where multiple computers are
connected. A computer network which has multiple OS
running is called server. All computers connected to
servers are called workstations or terminals.
The multiple users send the request to server and server
send back the result to each user after processing.
User-1 User-3
SERVER

User-2 User-4

Fig: Multiuser os
Ex:
• MS-windows NT server
• MS-Windows 2000 server
• MS-Windows 2003 server
• Novel Network
• Linux
• Unix
Written By: ATUL KUMAR Page 5
 OPERATING SYSTEM

➢ Function of OS:
• CPU management/processor
• Process management
• Memory management
• Disk management
• I/O management

➢ Evolution of OS: An OS may process the task serially or

concurrently. Based on this concept, there are following 3-
evaluation of OS:
a). Serial processing
b). Batch processing
c). Multi-programming

i. Serial processing: A serial processing system where the

instruction and data are entered into the computer
serially. The process of development and preparation of a
program in such an environment is slow and cumbersome
due to serial processing and manual processing.

ii. Batch processing: In batch processing the utilization of

computer resources got improve. In this process, jobs of
similar nature are collected from time to time and entered
the computer as a batch/group. The computer then
processes these jobs one by one without any user
intervention (interruption).

Written By: ATUL KUMAR Page 6

 OPERATING SYSTEM

A small program called “resident monitor” resides in

memory which automatically does the job of sequence
from are task to another.
The resident monitor is developed through a language
called JCL (Job Control Language).
In batch processing, the following technique is used to
improve the system performance.
• Buffering
• Spooling
• DMA

✓ Buffering: Buffering is a method of overlapping input,

output, and processing of a single job. After the data has
been entered in CPU starts processing it. The input device
is then instructed to start the next input immediately. This
way input devices and CPU are then both busy. This way it
improves the system performance.

✓ Spooling (Simultaneous peripheral operation online): This

technique is more efficient than buffering.
Buffering overlaps I/O and processing of a single job
whereas spooling allows CPU to overlaps input of one job
with computation and output of another job. Therefore,
this approach is better than buffering.

✓ DMA (Direct Memory Access):- DMA is a memory chip

which directly moves the block of data from its own buffer
Written By: ATUL KUMAR Page 7
 OPERATING SYSTEM

to main memory without intervention by CPU. While CPU

is executing, DMA can transfer data between high speed
I/O device and main memory. This way it increases the
throughput and system performance.
▪ Throughput: - The amount of work completed in there
unit time.

iii. Multiprogramming: Multiprogramming offers a more

efficient approach to increase the system performance by
keeping CPU or I/O device busy all the time.
Multiprogramming approach allows to MONITER

load more than one programs (jobs) Program-1

into memory at a time.
Program-2
An OS picks one program from memory
and starts executing it. When the “
current executing program waits for
Program-n
some I/O then the CPU switch over to
another program for execution. This Fig: Memory layout in

way it kept the CPU busy all the time. multi-programming

➢ Operating system architecture:- An OS is large and

complex software which supports large no of functions.
Therefore, it is developed as a collection of modules where
each module performs a particular task. The design of OS
is referred to as OS architecture. OS architectures are: -
a) Layered structure approach
b) Kernel approach
Written By: ATUL KUMAR Page 8
 OPERATING SYSTEM

c) Virtual machine
d) Client server model

➢ Layered structure approach: An OS architecture based on

layered approach consists of no. of layers (Level) where
each layer built on the top of lower layers.
The bottom layer is the hardware, and the highest layer is
the user interface. The first OS designed on layered
approach “THE” operating system which consist of 6-layers
as shown below: -
The bottom layer-0 deals with Layer-5 User program

hardware. Layer-1 handles Layer-4 Buffering for I/O device

the allocation of jobs to CPU. Layer-3 Device driver
Layer-2 Memory management
The layer-2 handles the task in
Layer-1 CPU scheduling
memory management such as:
creating virtual memory, Layer-0 Hardware

Fig: Layered structure of THE OS

swapped in, swapped out etc.
Layer-3 is responsible for handling and running a specific
device connected to the system. And the next layer-4 does
the buffering of the I/O device the highest layer-5 user
program providing interface to user communicated with
system.
The layers are designed in such a way that it uses
operation and services of layer below it. The main
advantage of layer approach is the modularity that helps in
debagging and verification of system easily. The

Written By: ATUL KUMAR Page 9

 OPERATING SYSTEM

The disadvantage of layer approach is the definition of

new layers.
➢ Kernel approach: In this approach an OS is basically
divided into 2 parts: kernel and shell.
A kernel is a part of an OS which directly communicates
with hardware. The kernel performs the following
function:
I. To provides a mechanism for creating and deleting of
process.
II. To provides processor scheduling, memory
management and I/O management.
III. To provides mechanism for synchronization of
process.
The shell is a part of OS which acts as interface between
user and kernel. It accepts the command for user and
convey to kernel for execution.
UNIX operating system is designed on kernel approach as
shown below: -
USER PROGRAM
SHELL
KERNEL

HARDWAREE

Fig: UNIX operating system

Written By: ATUL KUMAR Page 10

 OPERATING SYSTEM

➢ Virtual machine: - Virtual machine is a concept which

creates illusion of a real machine. It is created by virtual
machine OS which makes a single real machine appear as
several machine in following figure.
CPU

Virtual Machine OS

CPU CPU CPU

OS1 OS2 OS3

Printer Reader Printer Reader Printer Reader

Virtual Disk Virtual Disk Virtual Disk

Fig: Virtual Machine

From the user point of view virtual machines can be made

to appear like a real machine can be internally different.
Another important acceptance is that each user can run
OS of his own choice.

➢ Client/Server model: This is the commonly used

architecture of OS. In this architecture, an OS is divided
into two parts as: “client process” and “server process”.

Written By: ATUL KUMAR Page 11

 OPERATING SYSTEM

The client process sets the request to server process and

the server process sends back the result to
Server process
client process.
Result Memory service
Client
In this model kernel handles process Kernel
File service
the communication between Service
Terminal service
client and server and also manages memory
Fig: Client/ Server Model
service, file service
Terminal service, memory
service etc.
➢ Type of Operating System:
1) Batch Process: - An OS which supports the batch
processenvironment is called batch OS.
Similar jobs are grouped and entered the system.
There are some disadvantages of batch OS: -
a. The time between job submission and job completion
is very high. In other words, the output is delayed.
b. The programmer can’t correct the error the moment
it occurs.
c. The jobs are processed in order of submission.

2) Multiprogramming OS: - Multiprogramming OS is more

sophisticated than batch OS. The multiprogramming OS
has the potential to improve system throughput by proper
utilization of resources.
Various form of multiprogramming OS are: -
a) Multitasking OS
b) Multiprocessing OS
Written By: ATUL KUMAR Page 12
 OPERATING SYSTEM

c) Multiuser OS/Network OS
d) Time sharing OS.
e) Real time OS
I. Multitasking OS: - It is a form of multiprogramming OS
that performs multiple tasks simultaneously. It allows to
reside multiple programs into memory. When one
program waits for some I/O operation, an OS submits
another program to CPU for execution. This way, a user
gets the illusion that his/her multiple tasks are being
completed. Hence, it is also called serial multitasking and
it is different from multiprocessing.

II. Multiprocessing OS:- It is a type of multiprogramming OS

which supports parallel processing where more than one
process is executed concurrently by the multiple
computational unit (ALU). Hence, multiprocessing OS is
very fast as compared to multitasking OS.
Multitasking OS is based on single CPU whereas
multiprocessing is based on multiple CPU. This type of OS
is used in fast and complex systems like whether
forecasting, email processing, export systems, artificial
intelligence etc.

III. Multiuser OS/Network OS:- It is a type of

multiprogramming OS which has the ability to handle the
request of multiple users. It allows simultaneous access to
a computer system called server through two or more
Written By: ATUL KUMAR Page 13
 OPERATING SYSTEM

terminals. It is used in network environment and called

Network Operating System (NOS). A network OS provides
many capabilities like:
a) Allow users to access the various resources of
thenetwork host.
b) Controlling access so that only authorized users
canaccess the network resources.
c) Provides up to the minute network documentation
online.

IV. Time sharing OS: In the time-sharing OS allows many user

to use a particular computer system at the same time.
The CPU time is shared among multiple users. It is a form
of multiprogramming OS in which CPU switches rapidly
from one user to another user in a user-1

very small fraction of time (Time user-2 user-6

slot) and this way each user is switch

CPU

given the impression that his task user-3 user-5

is only being processed by the user-4

computer. Fig: Time sharing OS

The main advantage of time-sharing OS is that if reduce

the CPU idle time and provides quick response.

V. Real time OS: A real time system is defined as a data

processing system in which the data processing and
response time is very fast. The time taken to respond and
input, and display of output is called response time.
Written By: ATUL KUMAR Page 14
 OPERATING SYSTEM

The response time is very less as compared to online

processing.
It is used where fast response is needed. for example,
industrial control system, scientific experiment etc.
The major drawback of real time systems is that there is a
problem of data security and low volume of data
processing.

3) Distributed Operating System:- The distributed OS use

multiple CPU for executing users program. The data
processing jobs are distributed among the processors.
The use of multiple processors is invisible to the user, i.e.,
the user views the system as a uniprocessor but not as a
collection of different machines.
The user’s data may get processed on any CPU, but the
user is not aware of where programmers are being run or
where their files are stored.
The distributed system is more reliable than uniprocessor
based on system. Another advantage of the distributed
system is the incremental growth means more processor
added will give the more processing power.

❖ Process and process scheduling:- A program in running

state is called process. In other words, a program is in
passive state whereas process is an active state.

Written By: ATUL KUMAR Page 15

 OPERATING SYSTEM

❖ Process hierarchy: - An OS creates and kill processes.

When a process is created, it creates another process
which intern creates some more processes and so on and
this form process tree or hierarchy.
❖ Process states: - The lifetime of process is divided into
various stages called states. In other words, the state of a
process’s changes during its execution, each process may
be in one of the following states:
a. New: - The process has been created
b. Ready: The process waiting for the allocation to a CPU
(processor) for execution.
c. Running state: - The process being executed by CPU is
said to be running state.
d. Suspended/Waiting state: The process waiting for
some I/O and its execution is temporarily
paused/stops is said to be in suspended/waiting
states.
e. Terminate: A process is said to be in terminate state
when its execution is finished/completed.
New Ready Running Terminate

Suspended

Fig: Process state

/*Process Control Block(PCB*/

An OS groups all information that it needs about a
particular process in a data structure called Process
Written By: ATUL KUMAR Page 16
 OPERATING SYSTEM

Control Block (PCB). When a process is created, the OS

creates a corresponding PCB and when the process
terminated, the PCB also get released from memory.
A PCB contains the following piece of information about a
particular process.
✓ Process state: - Each process may be in states like
new,ready, running, waiting, terminate.
✓ Process number: - Each process is identified by a
uniquenumber called process id.
✓ Program counter (PC): It indicates the address of
instruction to be executed next.
✓ I/O- status information: It includes the information
about the I/O device allocated to a process.

/*Process Scheduling*/
Scheduling refers to the set of policies and mechanism
supported by OS that determines the order in which the
process/jobs/task will be completed. It is one of the main
functions of an OS. All computer resources are scheduled
before use.
The part of OS (Operating System) which performs the job
of scheduling is called scheduler. A scheduler selects an
available process from a set of processes for execution by
CPU. i.e., the selection of the process is carried out by the
scheduler. The main objective of scheduling is to increase
the CPU utilization and improve the overall efficiency of
the computer.
Written By: ATUL KUMAR Page 17
 OPERATING SYSTEM

➢ Type of scheduler: There are 3 types of schedulers.

i. Long term scheduler
ii. Medium term of scheduler
iii. Sort term scheduler

1. Long term scheduler: - It is also called job scheduler. The

long-term scheduler selects the process from mass storage
device and loads them into memory in a ready queue.
Long Term Short Term
Ready queue CPU Terminate
Scheduler Scheduler

Medium Term
Suspended Mode
Scheduler

Fig: Different scheduler task

2. Medium term scheduler: - The medium-term scheduler
isa scheduler which makes the suspended process to ready
state by loading into ready queue once the suspended
criteria is fulfilled.
3. Short term scheduler: - It is also called CPU scheduler or
dispatcher. It selects the process from ready queue and
admits to CPU for immediate processing.
❖ Various type of scheduling algorithm: -
a. First-Come First Serve (FCFS) Scheduling
b. Shortest-Job-First (SJF) Scheduling
c. Round-Robin (RR) Scheduling
d. Priority Based Scheduling

Written By: ATUL KUMAR Page 18

 OPERATING SYSTEM

e. Multi-Level Queue (MLQ) Scheduling

1. FCFS Scheduling: - It is a non-pre-emptive type of
scheduling algorithm.
This scheduling algorithm processes the job in the order of
their arrival i.e., the job which arrives first will be
processed/ executed first. It is based on FIFO (First-In-First-
Out) concept. A job has longer processed time than
another job has wait for longer. Hence, it has poor
performance.

Q1. Find the waiting time and turnaround time for the
following process under FCFS Scheduling.
Process Execution time
P1 5
P2 7
P3 9
P4 4
The process arrives in the order as P1→P2→P3→P4
➢ Process Waiting time Turnaround time
P1 0 5
P2 5 12
P3 12 21
P4 21 25

Avg. Waiting Time= (0+5+12+21)/4

=38/4
=9.5 unit
Written By: ATUL KUMAR Page 19
 OPERATING SYSTEM

2. SJF-Scheduling: - It is a non-pre-emptive type of

scheduling algorithm. In this scheduling algorithm, jobs are
processed based on the shortest executing time. The job
with equal execution time will be processed based on FCFS
algorithm.

Q1: Find the waiting time and turnaround time for

following process under SJF.
Process Execution time
P1 5
P2 7
P3 9
P4 4
The process arrives in the order as P1→P2→P3→P4
compare FCFS and SJF Who is best.
➢
Process Waiting time Turnaround time
P1 4 9
P2 9 16
P3 16 25
P4 0 4

Avg. waiting time= 29/4

=7.25 unit
Conclusion: - Since, Avg. waiting of SJF is less than the
avg.time in FCFS and therefore, SJF is more efficient.
3. Round Robin (RR) Scheduling:- It is also called Context

Written By: ATUL KUMAR Page 20

 OPERATING SYSTEM
Switching or Time Sharing scheduling. It is a pre-emptive
type of scheduling algorithm. The CPU time is divided into
time slices. Each process is allocated small and equal time
slices and CPU switched from one process to another
process in the given time slice.
If a process requires more time slices, then it waits for
next time slice. In this scheduling algorithm each user gets
impression that his/her only being processed.

Q1: Calculate waiting time and turnaround time for the

following process under RR scheduling (Time Slice: 5unit)
Process Execution Time
P1 25
P2 5
P3 5
Arrival order: P1→P2→P3
➢ 25 5 5

P1 P2 P3
0 5 X 10 X

Process Waiting Time Turnaround Time

P1 10 35
P2 5 10
P3 10 15
Avg. Waiting Time=25/3
=8.33 Unit
4. Priority Based Scheduling: - It is a non-preemptive type of
scheduling algorithm. In this algorithm each process is
assigned a priority (A number indicating level/precedence).

Written By: ATUL KUMAR Page 21

 OPERATING SYSTEM

A process is allocated to CPU for execution based on the priority

i.e., a process with higher priority is allocated to CPU before a
process will lower priority.
The process having the same priority is executed on FCFS
basis.
The process having the same priority is executed on FCFS
basis.
The major drawback of priority-based scheduling is the
infinite block of a low priority process this is called
starvation. A process is ready to run and waiting for CPU
but in the min time a process with higher priority comes in
ready queue then waiting process will not be submitted to
CPU. This infinite blocking of process of low priority is
solved through aging priority.
Aging priority is a technique of increasing the priority of
process that was it in the system for long time.

Q1: Calculated waiting time and turnaround time for the

following process under priority-based scheduling.
Process Execution Time Priority
P1 10 3
P2 1 1
P3 2 4
P4 2 5
P5 5 2
Process arrival order: P1→P2→P3→P4→P5

Written By: ATUL KUMAR Page 22

 OPERATING SYSTEM

➢ Process Waiting Time Turnaround Time

P1 6 16
P2 0 1
P3 16 18
P4 18 19
P5 1 6
Avg. Waiting Time=41/5
=8.1Unit
1. Multi-Level Queue (MLQ) Scheduling: - In this scheduling
algorithm, processes are classified into different groups.
for example, the interactive (foreground) process and
batch process (background). These 2 types of process have
different response time and so they are scheduled in
different manner. This scheduling algorithm partitions the
ready queue into system processes, interactive process
and batch process that create 3 ready queues as shown
below.
System Priority Based

Process Scheduling

Interactive R.R Scheduling

Switch CPU
Process

Batch FCFS

Process Scheduling

Written By: ATUL KUMAR Page 23

 OPERATING SYSTEM

/*Inter-Process Communication*/
In multiprogramming environment multiple processes are
executed concurrently (parallel). These concurrent
processes also communicate with each other which are
called inter-process communication.
The process of synchronization is also needed in case of
inter-process communication. Processes are executed with
very high speed and therefore 1-process must perform
some tasks before other process deletes it.
Synchronization can be viewed as setup constraints on the
ordering of the events.
Following are the some of the technique use in process
synchronization in multi programming environment: -

1. Mutual Exclusion: The processes that are working

together often share some common storage may be in
main-memory or it may be a shared file. Each process has
a segment of code called “Critical section” which accesses
shared memory or files. The key issue is to restrict more
than one process from reading and writing the shared data
at the same time.
Mutual exclusion is a technique that makes sure that if one
process is executing a critical section and accessing the
shared data then the other process will be excluded from
doing the same thing. The mutual exclusion needs to be
enforced only when processes accessed shared data.

Written By: ATUL KUMAR Page 24

 OPERATING SYSTEM

There is one drawback of mutual exclusion is that it

doesn’t handle a situation when critical section of multiple
processes gets executed at the same time.
Following are the 4-conditions to handle the problem of
critical section: -
a) No processes may at the sometime enter critical
section.
b) No assumptions are made about relative speed of
processes.
c) No processes outside the critical section should block
other processes.
d) No processes should wait long to enter its critical
section.
2. Semaphore: - The problem encountered in mutual
exclusion is overcome by a synchronization tool called
semaphore which was proposed by Dijkestra in 1965.
A semaphore is a variable which accepts nonnegative
integer value, and it is manipulated through the operations
“Wait” and “Signal”. Each process ensures the integrity of
its critical section by opening it with a “wait” operation
and closing with a “signal operation. This way any number
of con-current processes might share the resources
provided each of these processes use wait and signal
operation. A semaphore called binary semaphore (BSEM)
can contain values of 0 and 1. 0 indicates wait and 1
indicates signal. The signal (1) indicates that the resource
isavailable.
Written By: ATUL KUMAR Page 25
 OPERATING SYSTEM

❖ Dead lock: - Dead lock is a situation where a set of

processes are blocked because each processes is holding a
resource and waiting for another resource which is
acquired by some other processes.
The following diagram shows the situation of dead lock.
Resource-1

Process-1 Process-2

Resource-2

In the above diagram, process 1 is holding resource 1 and

waiting for resource 2 which is acquired by process 2 and
process 2 is waiting for resource 1. This situation leads to a
deadlock where execution of both processes is blocked.
Deadlock can arrive with following 4-condition -
a) Mutual exclusion: - Only one process at a time can
use resources.
b) Hold and wait: - A process is holding one resource
and waiting for another resource which is currently
held by another process.
c) No pre-emption: - A resource can’t be taken from a
process unless the process release the resource, in
other words the resources previous granted can’t be
precisely taken away from a process.

Written By: ATUL KUMAR Page 26

 OPERATING SYSTEM

d) Circular wait condition: - A set of process are

waitingfor each other for a resource in circular form.

➢ Handling dead lock: - There are 4 strategies used for

handling the dead lock situation.
a) Dead lock prevention or avoidance: - The idea is
tonot let the system enter deadlock state.
b) Detection and recovery: - Detect the situation of
deadlock and then do pre-emption to handle it.
c) Ignore the problem all together: - If the deadlock
is very rare then let it happen and reboot the
system.
d) Prevention by negating one of the 4-nesseary
conditions.

/*Memory management*/
Memory management is mainly concerned with allocation
of main memory to requested process. No process can
even run before a certain amount of memory is allocated
to it. Hence, organization and management of main
memory has been one of the most important factors to be
considered in design of OS. The overall resource utilization
and other performance criteria of a computer system are
mostly affected by the performance of the memory
management module.
Two important feature of memory management function
are protection and sharing.
Written By: ATUL KUMAR Page 27
 OPERATING SYSTEM

An active process should never attempt to access the

content of another process or destroy it. Apart from it,
memory management scheme must support sharing of
common data.
➢ Single process monitor: - In this memory management,
thememory is divided into 2-sections:
a.) Section for OS
b.) Section for user Program
Lower part Operation

Of Memory System

User

Program

Fig: Memory layout of single process monitor

In this memory management approach, an OS keeps track

of the 1st and the last location available for allocation of
user program.
When one program is completed and terminated, the OS
may load another program for execution i.e., only one
program is loaded at a time into memory. This type of
memory management scheme is commonly used in single
process OS such as MS DOS (Microsoft Disk Operating
System).
Sharing data in a single process environment doesn’t make
much sense because only one process resides in the main
memory at a time. Protection is hardly supported since,
only one program is residing at a time in memory and so it.
Written By: ATUL KUMAR Page 28
 OPERATING SYSTEM

may not occupy the whole memory. Therefore, memory is

under-utilized, and CPU will be sitting idle when a running
program requires some I/O operation.
/*Multiprogramming with fixed partition /*
In multiprogramming environment, several programs
reside in main memory at a time and the CPU passes its
control rapidly between these programs. One way to
support multiprogramming is to divide the main memory
into several sections/partitions where each partition is
allocated to a single process. Depending upon how and
when partitions are created,
There are two types of memory partitioning. These are
following: - OS
/////(Free)
P1
a) Static partitioning P2
P3
b) Dynamic partitioning /////(Free)

a) Static partitioning: - In static partitioning, the memory is

divided into number of partitions and its size is made in
the beginning itself which remain fixed thereafter.
b) Dynamic partitioning: - In dynamic partitioning the
memory is divided into a number of partition, but the
number of partitions and its size is decided at run time
by the OS. Each partition will store a single program for
execution. The no. of programs that can reside in
memory is bounded by the number of partitions. When
the program terminates the partition is free for another
program.

Written By: ATUL KUMAR Page 29

 OPERATING SYSTEM

➢ PDT (Partition Description Table):- When partitions are

defined/ created, an OS keeps track of the status of
memory partition into a data structure called “PDT”. In
PDT, the OS maintains the information like starting
address of partition, its size and status as free or allocated.
Partition No. Starting Address Size of partition Status
1 OK 200K Allocated
2 200k 200k Free
3 400k 200k Allocated
4 600k 300k Allocated
5 900k 100k Free
6 1000k 100k Free

Fig: Partition Description Table

➢ Partition allocation method: - There are 2-most commonly

techniques used to allocated free partition to ready
process. These are the following.
a) First Fit
b) Best Fit
a) First fit: - In this partition allocation method, an OS
allocates the first free partition to a requested process
whichis large enough to accommodate the process.

b) Best fit: - In this partition allocation method, an OS

searches the entire PDT and allocated the free partition to
the requested process which is smallest to fit requirement.
Conclusion: - The first fit method is faster than best fit
method but in case of first fit lots of memory get wasted

Written By: ATUL KUMAR Page 30

 OPERATING SYSTEM

whereas the best fit method achieve higher memory

utilization.
Q1: Consider the following given PDT and process to be
loaded show the allocation of partition to process in
first fit aswell as in best fit.
Partition No Address Size Status Process Size
1 100 100k Free P1 212k
2 200 500k Free P2 417k
3 700 200k Free P3 112k
P4 426k
4 900 300k Free
5 1200 600k Free Fig: Process to be loaded

Fig: PDT

➔ Partition allocation in first fit algorithm: -

Process Size Allocation
P1 212k 500k
P2 417k 600k
P3 112k 200k
P4 426k X(Can’t be loaded)
➔ Partition allocation in best fit algorithm: -
Process Size Allocation
P1 212k 300k
P2 417k 500k
P3 112k 200k
P4 426k 600k

In the first fit algorithm, the process P4 does not get

loaded as the available partition is not able to fit the
process. On the other hand, in the case of the best fit
algorithm, all process gets loaded in memory. Hence, the
best fit algorithm utilizes memory efficiently rather first fit.

Written By: ATUL KUMAR Page 31

 OPERATING SYSTEM

➢ Multiprogramming with dynamic partition: - The main

drawback with fixed size partition is the wastage of
memory when the programs are smaller than partitions.
This type of wastage of memory is also called internal
fragmentation.
A memory management approach known as dynamic
partitions or variable partitions which creates the partition
dynamically to meet the requirement of each process.
Compared to multiprogramming with fixed partition, in
multiprogramming with dynamic partition, the size,
location and the nod of partition very dynamically as
process are created and terminated whereas they are fixed
in fixed size partition approach.
In this approach the memory manager allocated partition
to the requesting process until all the physical memory is
exhausted. Therefore, memory utilization gets improved.
The main problem in this approach is the external
fragmentation i.e., even if there is enough memory to
occupy the process, it can’t satisfy a request due to non-
contiguous memory allocation. External fragmentation
happens when the storage is fragmented into a smaller
number of holes (free space).

OS OS
100k P1 Free //100k//
300k P2 300k
50k P3 Free //50k//
200k P4 200k
Fig: External Fragmentation

Written By: ATUL KUMAR Page 32

 OPERATING SYSTEM

➢ Compaction: - In dynamic partition external fragmentation

occurred as the process terminated. This external
fragmentation problem is solved through compaction.
Compaction is a process of combining all free holes into a
large block by pushing all the process downwards as far as
possible. Compaction is usually not done because it
consumes lots of CPU time.
The following figure shows the compaction of memory.

OS OS
P1 P1(100k)
//50k// Free P3(200k)
P3(200k) Compaction P5(300k)
//100k// Free
P5(300k) //180// Free
//30k// Free

Fig: Compaction of memory

➔ Advantages: -
I. Memory utilization is better than fixed size
partitioning since partition is created
according to the size of process.
II. Protection and sharing in fixed and dynamic
partition are similar.
III. When the process is larger than the free
partition then OS expends the free area by
combining the adjacent free area and moving
the process into it.

Written By: ATUL KUMAR Page 33

 OPERATING SYSTEM

→ Disadvantages: -
I. Dynamic memory management requires lots
of OS space, time, and complex memory
management algorithms.
II. Compaction time is very high.

/*Paging*/
Paging is a memory management scheme/technique which
allows a process to be stored in memory in non-contiguous
partition. This way it solves the problem of external
fragmentation. The program is divided into logical memory
space or virtual address space called pages. Similarly,
physical memory is also divided into the fixed/same size
blocks called frames or page frames. When a process
needs to be executed then its pages are loaded into any
frames of physical memory from the disk.
The program generated addresses are called virtual
address and the address where actually the page will be
stored is called physical address. Hence, address mapping
is required to convert the logical address into physical
address.
The logical address consists of page no. and offset
whereas physical address consists of base address and
offset.
The following diagram shows the address mapping in the
paging system.

Written By: ATUL KUMAR Page 34

 OPERATING SYSTEM

Physical memory
Physical address
Virtual address
CPU Page no. Offset Basic addres Offset

Base address

Page Map Table (PMT)

Fig: Address mapping in paging system

The Paging system uses tables called Page Map Table

(PMT) to store base addresses which consist of page
number and offset.
➢ Segmentation: - Segmentation is a memory management
scheme which supports user’s/programmer’s view of
memory. In this memory management technique, the
process is divided into variable size segments and loaded
to the logical memory address space. The logical address
space is the collection of variable size segment where each
segment has its length and name.
The address specified by the user contains segment name
and the offset and segments are not called segment no.
Segment avoids internal fragmentation but leads to
external fragmentation.
Segmentation uses segment table in address mapping
system where the virtual address is converted into physical

Written By: ATUL KUMAR Page 35

 OPERATING SYSTEM
address. The segment table contains segment no., base address,
and the size of segment.
The following diagram shows the address mapping in
segment system.
Segment 3
Virtual
Segment no. Offset Segment 2
CPU Address
Segment 1
Base address Size
Segment 0
1 4000 300
Physical map
2 6000 400
3 --
4 --
Segment
Fig: Virtual to physical memory mapping in segment system

➢ Demand Paging: - Demand paging is another memory

management scheme used in a paging system when
memory capacity is low. In demand paging, the pages are
loaded into memory only on demand and not in advance.
It is like a paging system with a difference that instead of
swapping the entire program into memory, only those
pages are swapped which are required currently by the
system.
When a program tries to access a page which was not
swapped in memory then it is called page fault trap. When
a Page fault trap occur then valid page is again loaded into
memory. An OS handles the page fault in the following
ways: -
i) The OS checks whether the memory reference for
the missing page was valid or not.
Written By: ATUL KUMAR Page 36
 OPERATING SYSTEM

ii) If the memory reference is valid but the page is

missing, then the process of bringing a page into
physical memory starts.
iii) Free memory location is identified to bring the
missing page.
iv) The page is read from disk and loaded into
memory location.
v) Page map table is updated with the process/page
brought in memory.
vi) Record the instruction which was intercepted
due to missing page.

/*Disk management*/
Disk management is an important function of an OS. The
performance of a computer system also largely depends
upon how fast a disk request is serviced. In a
multiprogramming environment, many processes may
generate request for reading and writing disk record. To
service a request, the disk system requires moving the
head at desired track then wait for latency and finally
transfer the data. When more than track is to be serviced
to different processes then the order in which track will be
serviced to process depends upon the disk scheduling
algorithm.
➢ Disk scheduling algorithm: Following are the various
scheduling algorithm for disk services: -
i. FCFS Scheduling
Written By: ATUL KUMAR Page 37
 OPERATING SYSTEM

ii. SSTF Scheduling

iii. SCAN Scheduling

1. FCFS (First Come First Served Scheduling): - This is the

simplest form of disk scheduling in which the first request
to arrive is the first one to be serviced. It is easy to
implemented but doesn’t provide the best services.
Consider the following track read under FCFS
scheduling.
100, 200, 50, 150, 25, 155, 70 and 85
Starting track to read in 100 and last to read in 85, let the
head is initially positioned at track 50 find the total head
movement.
P 15 25 50 70 85 100 150 155 200

Written By: ATUL KUMAR Page 38

 OPERATING SYSTEM

Total head movement: - (100-50)+(200-100)+(200-50)+

(150-50)+(150-25)+(155-25)+(155-70)+(85-70)
=> 50+100+150+100+125+300+85+15= 755 Ans

2. SSTF (Shortest Seek Time First): - In SSTF scheduling,

priority is given to those processes which need the
shortest seek time. In other words, the requested track
which is nearer to current head position will be serviced
before the tracks which are far away from the head
position.
One of the disadvantages of SSTF scheduling of SSTF
scheduling is that the inner-most and outermost tracks will
receive poor services as compared to need range tracks.

3. Scan scheduling: - In this scheduling algorithm, the read/

write head starts from one end and moves to another end
and services the requested (tracks). Which comes on the
path of head movement.
After reaching another end the disk head reverses its path
and continues the request which even comes in the path.
This way the head continues to oscillate from one end to
another end. Hence, this algorithm is also called “Elevator”
algorithm.
4. C-Scan scheduling (Circular): - In this scheduling algorithm,
the read write head moves from one end to another end
and services the tracks which come on the path of head
movement. After reaching to another end of disk, the
Written By: ATUL KUMAR Page 39
 OPERATING SYSTEM

read/write head reverses its direction to move to another

end. It doesn’t service the request while reversing the
direction.

5. Look scheduling: - This disk scheduling algorithm is like

Scan scheduling with a difference in that look scheduling
reverses its path of movement of read/write head after
services the last track instead of jumping to the last tracks
of disk.

6. C-Look scheduling: - This disk scheduling algorithm is like

C-Scan algorithm with a difference in that the read/ write
head reverses its movement of path after services the last
requested and doesn’t jump the last track of disk. Like C-
scan, it doesn’t provide services while reversing its path.

/*Disk space management method*/

The Operating System maintains a list of free disk space
tokeep track of all disk blocks which are not being used.
Whenever a file is to be created, the list of free disk space
is searched for and then allocated to the new file. The
amount of space allocated to the file is removed from
the free space list and when the file is deleted, the free
disk space is added to free space list. Following are the
two methods used to manage free disk blocks: -
i. Linked list
ii. Bitmap
Written By: ATUL KUMAR Page 40
 OPERATING SYSTEM

1. Linked-list: - In this method, all free disk blocks are linked

together by each free block pointing to the next free block.
There is an extra pointer pointing to the first free block of
the linked list.
Following diagram show the linked list method: -
0 6 12 18
1 7 13 19
2 8 14 20
3 9 15 21
4 10 16 22
5 11 17 23

Fig: Linked List of free disk block End

❖ Drawback of linked-list method: -

i. To reach a specific free block, traversing starts from
the beginning and hence, it takes substantial time.
ii. The pointer maintains that free disk block
requiresadditional disk/memory space.
iii. A single break in the link makes the disk block
inaccessible

2. Bit map: - In this disk space management method, binary

digits are used for indicating free blocks and allocated
blocks. The binary digits 0 is used for marking free block
whereas 1 is used for marking allocated block.

Written By: ATUL KUMAR Page 41

 OPERATING SYSTEM

0 6 12 18
1 Free 7 Free 13 19
2 8 14 Free 20
3 Free 9 15 21
Free 4 10 16 22
5 Free 11 17 Free 23

Fig: Allocated and free block disk

In this above figure blocks 4, 7, 9, 11, 13, 20 and 23 are

free block and rest of the block are allocated to file.
The bit map representation for the above shows free and
allocated block will be as follow:
111101101010101111110110
One of the main advantages of this method is that it is
simple and efficient to find the free block on the disk, but
the disadvantage is that it requires extra disk space to
store bitmap.

/*Disk allocation method*/

There are 2-popular methods for allocation free disk-block
to file these are: -
a) Contiguous allocation
b) Non-contiguous allocation

1. Contiguous allocation: In this method, files are assigned

to contiguous area of secondary storage. A user
specifies the size of area in advance needed to hold a
file and if the desired amount of contiguous space is not
available the file can’t be created.
Written By: ATUL KUMAR Page 42
 OPERATING SYSTEM

This method uses following to ways to allocated free

block: -
a) First fit
b) Best fit

I. First fit: - In this case, as soon as the first free block is

large enough to store the file encountered, it is
allocated.
II. Best fit: - In this case, the smallest block large enough
to accommodate the file is searched and then
allocated.
First fit is faster than best fit but best fit does the
efficient management of disk space. In the case of the
first fit lots of disk space is wasted.

2. Non-contiguous allocation: - In this method files are

allocated to non-contiguous free block. It uses the
following two techniques to keep track of non-
contiguous blocks.
a) Linked allocation
b) Indexed allocation
I. Linked allocation: - This method uses linked list for
maintaining the allocated non-contiguous block.
Because of non-contiguous allocation there is no
external disk fragmentation. It is simple and doesn’t
require disk compaction, but it doesn’t allow direct
access.
Written By: ATUL KUMAR Page 43
 OPERATING SYSTEM

Also, there is a problem of reliability and takes

extraspace for storing pointer.
0 4 8 12
1 5 9 13
2 6 10 14
3 7 11 15
File: Start From 3 End
Fig: Linked allocation of disk block

II. Indexed allocation: - In this method, the allocated

disk blocks are maintained by creating the index of
allocated block. Each indexed points to the allocated
block one of the major advantages of indexed
allocation is that the index allocation allows direct
accessing. The disadvantages of index allocation are
that it is more complex and takes more memory for
maintaining index. Index File
File
My file 3,5,7,10,15 0 4 8 12
1 5 9 13
2 6 10 14
3 7 11 15

Fig: Index allocation of disk block

THE END
THANK YOU
Written By: ATUL KUMAR Page 44