4.1.

Basic Resources Managed by Operating Systems

Introduction

Operating system is a collection of programs. It is a software that supports computer's basic

functions. It acts as an interface between applications and computer hardware as shown below in

the figure 4.1. Examples of operating systems are Windows, Linux, MacOS, etc.

Operating system also does the task of managing computer hardware. Operating system

manages hardware by creating a uniform set of functions or operations that can be performed on

various classes of devices (for instance read x bytes at address y from hard drive of any type,

where x is the number of bytes read). These general functions rely on a layer of drivers that

provide specific means to carry out operations on specific hardware. Drivers usually are provided

by the manufacturers, but the OS must also provide a way to load and use drivers. OS must

detect the device and select an appropriate driver if several of them are available.

In the figure 4.2 below, OS interface is a set of commands or actions that an operating system

can perform and a way for a program or person to activate them.

Operating system is the core software component of a computer. It performs many functions and

acts as an interface between computer and the outside world. A computer consists of several

components including monitor, keyboard, mouse and other parts. Operating system provides an

interface to these parts through 'drivers'. This is why, when sometimes we install a new printer or

other piece of hardware, our system will ask us to install more software called driver. A brief on

device drivers is given in section 4.2.4, Device Management.

Operating system basically manages all internal components of a computer. It was developed to

better use the computer system.

4.1. Basic Resources Managed By Operating System

Following are the basic resources that are managed by operating system:
  Memory space (Memory management)

 CPU time (Process management)

 Disk space (File system management)

 Input-output device (Device management)

4.1.a. Basics Of Memory Management

Any program needs to be loaded in memory (RAM) before it can actually execute. Random

Access Memory (RAM) is a type of computer memory that can be accessed randomly, i.e. any

byte of memory can be accessed without touching or traversing through the preceding bytes.

These days, RAM is available with various storage capabilities (256MB, 512MB, 1GB, 2GB,

4GB, 6GB and 8GB sizes).

Memory management deals with managing computer's primary memory. It decides the amount

of memory to be allocated to a particular program. It also keeps track of free or unallocated as

well as allocated memory. Memory management applies only to RAM.

The sole motto of memory management done by operating system is to utilize the available

memory effectively. In this process we should try to minimize fragmentation.

i. Fragmentation

Fragmentation occurs when there are many free small blocks in memory that are too small to

satisfy any request. In computer storage, fragmentation is a phenomenon in which storage space

is used inefficiently, resulting in reduced capacity as well as performance. Fragmentation also

leads to wastage of storage space. The term also refers to the wasted space itself.

There are three different types of fragmentation:

 External fragmentation

 Internal fragmentation

 Data fragmentation
It can be present in isolation or conjunction. When a computer program is finished with a

partition, it can put the partition back to the computer. The size and amount of time for which a

partition is held by a program varies. During its life span, a computer program can request and

free many partitions of memory.

When a program starts, there are long and contiguous free memory areas. Over the time and

with use, these long contiguous regions become fragmented into smaller contiguous areas.

Eventually, it may become impossible for the program to request large partitions of memory.

Internal Fragmentation

Internal fragmentation is the space wasted inside the allocated memory block because of the

restriction on allowed sizes of allocated blocks. Allocated memory can be slightly larger than

requested memory. The size difference is the memory internal to partition but would not be used.

Internal fragmentation is difficult to reclaim. Design change is the best way to remove internal

fragmentation. For example, in dynamic memory allocation, memory pools cut internal

fragmentation by spreading the space overhead over a larger number of objects.

External Fragmentation

External fragmentation happens when a dynamic memory allocation algorithm allocates some

memory and a small piece is left over which cannot be used effectively. The amount of usable

memory reduces drastically, if too much external fragmentation occurs. The total memory space

available is enough to satisfy a request but is not contiguous and so, is wasted. Here the jargon

'external' is used to depict that the storage that is further unusable is outside the allocated

regions.

For example, consider a situation wherein a program allocates three continuous blocks of

memory and then frees the middle one. The memory allocator can use this free block of memory

for future allocations. However, this free block cannot be used for allocation if the memory to be

allocated is larger in size than the free block. External fragmentation also occurs in file systems

as many files of different sizes are created, size changed and deleted. The effect is even worse if

a file which is divided into many small pieces is deleted, because this leaves almost equal small

regions of free space.

Data Fragmentation
Data fragmentation occurs when a collection of data in memory is broken up into many pieces

that are not close enough. It is typically the result of attempting to insert a large object into

storage that has already suffered external fragmentation.

ii. Memory Management Techniques

 Single Contiguous Allocation

 Partitioned Memory Management Scheme (refer Glossary for scheme definition)

 Paged Memory Management

 Segmented Memory Management

Single Contiguous Allocation

Single Contiguous Allocation is the simplest form of memory management. In this type of

memory management, all memory (with an exception of small portion reserved for running the

operating system) is available for a single program to run. So only one program is loaded in all

available memory and so generally the rest of memory is wasted. Simplest example of this type

of memory management is MS-DOS. Advantage of single contiguous memory allocation is, it

supports fast sequential and direct access. Provides good performance and the number of disk

seek required is minimal. Disadvantage of single contiguous memory allocation is fragmentation.

Partitioned Memory Management Scheme

This type of memory management divides primary memory into multiple memory partitions,

usually contiguous areas of memory. Memory management here consists of allocating a partition

to a job or program when it starts and unallocating it, when job or program ends. A typical

scenario of partitioned memory management is depicted in the figure 4.4 below. This type of

memory management also needs some hardware support.

The figure 4.4 below shows multiple memory partitions. Each partition runs a process (Refer

Glossary). Process 1, Process 2, etc. are the processes allocated to a particular partition in

memory. Once the process is completed, partition gets empty. A is the base register address.
Base register contains the lowest memory address a process may refer to. We can get length of

a partition from Bounds register. Bounds register stores upper and lower bounds on addresses in

a time sharing environment. Time sharing environment refers to concurrent use of computer by

more than one user – users share the computer’s time. Time sharing is the term used

synonymously with multi-user.

Partitioned memory management is further divided in to 2 types of partitioning schemes:

 Fixed memory partitioning (Static partitioning)

Main memory is divided into a number of static partitions when a computer system starts. A

process may be loaded into a partition of equal or greater size. Advantage of such type of

partitioning is that it is simple to implement and has a little operating system overhead.

Disadvantages are inefficient use of memory due to internal fragmentation. Also maximum

number of active processes is fixed.

 Variable memory partitioning (Dynamic partitioning)

Such partitions are created dynamically so that each process is loaded into a partition of exactly

the same size as that process. Advantage of such type of partitioning is that there is no internal

fragmentation. Also main memory is used more efficiently. Disadvantage is inefficient use of

processor due to the need for compaction to counter external fragmentation.

For further details on types of partitioning please refer below link:

http://www.csbdu.in/econtent/Operating System/unit3.pdf

Paged Memory Management

This type of memory management divides computer's primary memory into fixed-size units

known as page frames. The program's address space is divided into pages of the same size.

Usually with this type of memory management, each job runs in its own address space.

Advantage of paged memory management is that there is no external fragmentation. But on the

contrary there is a small amount of internal fragmentation.

Below is the Illustration representing paged memory management. It depicts how logical pages

address in memory can be mapped to physical address. The operating system uses base

address as a measure to find addresses. Base address means starting address of a particular

memory block. According to the program written, CPU generates address. In this case, address

generated by CPU is called logical address. This address is added to base address so that it

forms physical address. To translate a logical address into corresponding physical address, we

need to divide the logical address into page number and offset. Offset refers to a value being

added to base address to produce a second address.

For example, if B represents address 200, then the expressions, B+10 would signify the address

210. 10 in expression is the offset. To specify addresses using an offset is called relative

addressing because the resulting address is relative to some other point. Offset is also known as

displacement. In figure below, p represents page number and d is the offset. We use page

number as index to page table and the entry gives corresponding frame number. 'f' in page table

represents frame number. Frame number is concatenated with the offset to get a corresponding

physical address.

Segmented Memory Management

Segmented memory means, the division of computer's primary memory into segments or

sections as shown in figure 4.6. In this type of memory management, there is reference to a

memory location that includes a value . This value identifies a segment and an offset within that

segment. We may create different segments for different program modules, or for different types

of memory usage such as code and data segments. Certain memory segments may even be

shared between programs. Segmented memory does not provide user's program with a linear

and contiguous (or continuous) address space. Linear address space is a memory addressing

scheme used in processors where the whole memory can be accessed using a single address

that fits in a single address or instruction.

Segments are areas of memory that usually correspond to a logical grouping of information such

as code procedure or a data array (Refer Glossary). Segmentation allows better access

protection than other memory management schemes because memory references are relative to
a specific segment and hardware will not permit the application to reference memory not defined

for that segment. It is possible to implement segmentation with or without paging.

Segmentation With Paging has the following advantages

1. Paging eliminates external fragmentation. As a result it provides efficient use of main

memory.

2. Segmentation which is visible to the programmer, includes strength of paging. It has the

ability to handle growing data structures, support for sharing, and modularity, and protection.

To combine fragmentation and paging, a user's address space is broken into a number of

segments. Each of these segments in turn is broken into fixed-size pages. If size of a segment is

less than a page in length, then the segment occupies just one page.

4.1.b. Process Management

CPU has multiple processes running at a time. CPU always has certain programs scheduled in a

queue waiting for their turn to execute. These all programs must be scheduled by operating

system.

Operating system must allocate resources to processes, enable processes to exchange and

share information, enable synchronization among processes, and protect resources of each

process from other processes. Synchronization means the coordination of events or processes

so that a system operates in unison.

Process creation involves four principle events:

 System initialization or start up. When an OS is booted, typically several processes are
created.

 A process is run to execute a process creation system call. Most of the times, a running
process will issue system calls to create one or more new processes to help it do its job.
Creation of new processes is useful when the work to be done can be easily formulated as
several related, but otherwise independent interacting processes.
  Request to create a new process is issued by user. In interactive systems, users can start a
program by clicking or double clicking the icons or thumbnails or even by typing a command
on command prompt.

 Batch job initiation. Users can submit batch jobs to the system (most of the times,
remotely). When OS knows that it has the resources to run another job, a new process is
created and the next job is run from the input queue.

Several processes are created when an operating system boots. Some of them are foreground

processes while others are the background processes. Foreground processes interact with a

user and perform work for users. In multiprocessing environment, the process that accepts input

from the keyboard or other input device at a particular point in time is at times called the

foreground process. This means that any process that actively interacts with user is a foreground

process. Background processes are the processes which are not associated with a particular

user, but instead have some specific function to be performed. For example email notification is

an example of background processes. System clock displayed at the bottom right corner of

status bar in windows is another example of background process.

Some of the common reasons for process termination are:

 User logs off

 A service request to terminate is executed by process

 Error and fault conditions

 Normal completion

 Time limit exceeded

 Memory unavailable

 I/O failure

 Fatal error, etc.

Figure 4.7 below contains great deal of information.

 Let us consider a running process 'P' that issues an input-output request

o The process blocks.
o Thereafter, at certain point of time later, a disk interrupt occurs and the driver
detects that P's request is satisfied.
 o P is unblocked, i.e. the state of process P is changed from blocked to ready.
o Later, at some point of time, the operating system looks for a ready job to run and
picks the process/job P from the queue.

 A preemptive scheduler in figure 4.7 below has the dotted line 'Preempt'; where as a non-
preemptive scheduler doesn't.

 The number of processes change for two arcs namely, create and terminate.

 Suspend and resume are a part of medium term scheduling

o It is done on a bit longer time scale.
o It involves memory management as well.
o It is also known as two level scheduling.

There are 2 types of scheduling:

1. Preemptive scheduling:

In a computer, tasks are assigned priorities. At times it is necessary to run a certain task that has

high priority before another task (even if the task is in running state). Therefore, the running task

is interrupted for some time, put to either blocked or suspended state and resumed later, when

the priority task has finished its execution. This process is called preemptive scheduling.

2. Non preemptive scheduling:

In non-preemptive scheduling, a running task executes till it completes fully. It cannot be

interrupted. That means when a process enters the running state, it cannot be deleted from the

scheduler until it finishes its service time.

The figure 4.7 above shows various states of a process. Initially the process is created. Once

created, it goes to the ready state. Ready state means the process is ready to execute. It has

been loaded into the main memory and is waiting for execution on a CPU. There can be many

ready processes waiting for execution at a given point of time. A queue of processes which are

ready to execute gets created in memory. One of the processes from that queue is picked up for

execution and its state gets changed to running. A running process can be blocked. Blocked

state means a process is blocked on some event.

A process may be blocked due to various reasons such as when a particular process has

exhausted, the CPU time allocated to it, it is waiting for an event to occur. Blocked process can
either move to ready state or can move to suspended state. In systems that support virtual

memory, a process may be swapped out, that is, it would be removed from main memory and

would be placed in virtual memory by the mid-term scheduler. This is called suspended state of a

process. From here the process may be swapped back in to the ready state. Such state is called

ready suspended state. Process that are blocked may also be swapped out. Such a state of

process where a process is both swapped out and blocked is called blocked suspended state.

Suspended processes can be sent back to ready state only once they are released. This cycle

continues till a process finishes its execution i.e. terminated. A process may be terminated from

the running state by completing its execution or can be killed explicitly. In either of these cases,

we say that the process is terminated.

4.1.c. File System Management

Computer contains numerous files. They need to be organized in a proper way so that we can

keep track of those files. File retrieval should be easier as well as faster. File system

management helps us achieve this.

File system manager is used by the operating system to do file system management. File system

manager organizes and keeps track of all the files and directories on secondary storage media.

Figure 4.8 above shows a typical hierarchical file system structure. Operating system keeps track

of following tasks for providing efficient file management:

 It is able to identify numerous files by giving unique names to them.

 It maintains a list to keep track of exact file location.

 Provide fast and simple algorithms to write and read files in co-operation with device
manager.

 Grant and deny access rights on files to programs and users.

 Allocate and de-allocate files so that files can be processed in co-operation with process
manager.

 Provide programs and users with simple commands for handling files.
On storage medium, file gets saved in blocks or sectors. To access these files, file manager and

device manager works together. The device manager knows where to find each sector on disk,

but only file manager has a list telling in what sectors either file is stored. This list is called File

Allocation Table (FAT). FAT has been in use under DOS for a long time. Some of its alterations

are still used by Win95 and Win98.

Below are the different ways of allocating files:

 Contiguous file allocation:

In this type of file allocation, at the time of file creation, a single set of blocks is allocated to a file.

Each file is stored contiguously in sectors, one sector after another. The advantage is that the

File Allocation Table has a single entry for each file, indicating the start sector, the length, and

the name. Moreover, it is also easy to get a single block because its address can be easily

calculated. The disadvantage may be that it can be difficult to get a sufficiently large block of

contiguous blocks. Contiguous file allocation is now a days only used for tapes and recordable

CDs.

 Non-contiguous file allocation:

With this type of allocation, all blocks of a file can be distributed all over the storage medium. The

File Allocation Table (FAT) lists not only all files, but also has an entry for each sector the file

occupies. As all information is stored in the FAT, and there is no assumption on the distribution of

the file, this method of allocation, at times, is also known as FAT.

The advantage is that it is very easy to get a single block, because each block has its entry in the

FAT. Also, it is a very simple allocation method where no overhead is produced and no search

method for free blocks is needed. The disadvantage is that FAT can grow to an enormous size,

which can slow down the system. Compaction would be needed time to time. This type of file

allocation is used for disk allocation in MS-DOS.

 Chained allocation:

With chained file allocation only the first block of file gets an entry in the FAT. A block has sectors

in it. Of these, the first sector has got data as well as a pointer at its end that points to the next

sector to be accessed after it (or indicates that it was the last). That means each sector (if it is not
a last sector) has got a pointer at its end pointing to the next sector that should be accessed. The

advantage again is that the FAT has a single entry for each file, indicating position of the first

sector and file name. There is no need to store the files contiguously.

The disadvantage of chained allocation is that it takes much time to retrieve a single block

because that information is neither stored anywhere nor can it be calculated. If we want to

access a particular sector, all preceding sectors have to be read in order one after another. This

way, the information about location of the next block is identified. Unix i-node is an example of

this type of allocation. i-nodes are data structures (Refer Glossary) of a file system used to store

all the important properties of each file: owner's group id and user id, access permission on that

file, size of the file and more. Unix i-node is an operating system which has i-nodes as its data

structure.

 Indexed allocation:

With indexed file allocation also only the first block of file gets an entry in the FAT. In the first

sector, no data is stored. It has only pointers to where the file is on storage medium. That is why

the first block is known as the index block. It just contains the pointers or index to file location on

a storage medium. Here also the FAT will only contain a single entry for each file, indicating file

name and position of the first sector. It is easy to retrieve a single block because the information

about its location is stored in the first block.

The disadvantage of indexed allocation is that for each file an additional sector is needed to keep

track of the location of a file. A very small file also always occupies at least two blocks, where as

the data could have easily fit in one block. This results in secondary storage space wastage.

Indexed allocation is implemented in all UNIX systems. It is reliable as well as fast. Also, now-a-

days, the wastage of storage space does not matter much.

The main concern of file system management is to provide a strategy that lets the FAT not to

grow too large and that makes it possible to retrieve a special sector of file such that the storage

space is not wasted much.

4.1.d. Device Management

Operating system also deals with device management, specially input and output devices. In a

multi-user system, there are shared devices like printer, scanner, etc. Management of these

devices by sending them appropriate commands is the major task of operating system.

A software routine which knows how to deal with each device is called a “driver” and operating

system requires drivers for the peripherals attached to computer. When a new peripheral is

added, device driver is installed in to the operating system.

Operating system also deals with the access time of these devices. It helps make the device

access fast and in the most efficient way possible.

Major concern of operating system in device management is to prevent dead lock situation.

Dead Lock

Above diagram depicts a typical dead lock situation, that serially used devices can have. Dead

lock is a situation where each set of processes is waiting for an event that only other process in

the set can cause.

For example, in the above diagram, Process Task 1 requests a mutex or lock on object which is

already held by Process Task 2. In turn, Process Task 2 needs a resource to complete its

process which is already held by Process Task 1. So neither of the processes are releasing their

resources or mutex and waiting for another process to release a mutex so that they can get it to

complete its pending task. This situation would cause an indefinite wait for processes leading to

a situation known as dead lock.

Mutex is the short form for Mutual Exclusion Object. A mutex is a logical unit. It is a program

object that allows multiple program threads to share the same resource, but not simultaneously.

File access is an example of mutex. A mutex with unique name is created when the program

starts. Mutex must be locked by a thread that needs the resource. When data is no longer

needed, the mutex is set to unlock.

OS helps in dealing with dead lock situation up to a certain extent. Below are some of the

strategies for dealing with deadlock:

 Ostrich algorithm, i.e. ignoring the problem altogether. Ostrich algorithm is a strategy of
ignoring potential problems on the basis that they may be exceedingly rare. Ostrich
 algorithm assumes that ignoring the problem would be more cost-effective as compared to
allowing the problem to occur and then attempt for its prevention. It may occur very
infrequently. This can be followed if cost of detection/prevention is not worth the time and
cost spent.

 Detection and Recovery.

 Avoiding deadlock by careful resource allocation.

 Prevention of deadlock by structurally negating one of the four necessary conditions.

4.2. Resource Sharing Management

The primary task of operating system as discussed above is to keep track of resources like

memory, CPU utilization, storage device, and input output devices, to grant resource requests, to

mediate conflicting requests from different programs, etc.

4.2.1. Multiplexing

Multiplexing is a method by which multiple analog data signals/digital data streams are combined

into one signal over a shared medium. Multiplexing is used to share an expensive resource and

thus help reduce the cost.

Communication channel is used for transmission of a multiplexed signal. This channel may be a

physical transmission medium. Multiplexing basically works on the principle of division of signals.

It divides the capacity of the high level communication channel into several low level logical

channels. Each of these low level logical channel maps to each message signal or data stream

to be transferred. A reverse of this process is known as demultiplexing. Demultiplexing can

extract the original channels or signals on the receiver side.

A device that performs the task of multiplexing is called a multiplexer (MUX), and a device that

performs the task of demultiplexing is known as a demultiplexer (DEMUX).

Resource management basically includes multiplexing the resources in two ways:

Time Multiplexing

When a particular resource is time multiplexed, different programs or users get their turn to use

that resource one after another. Turns for resource usage are decided as per the predefined
operating system algorithm. Printer is one of the best examples of in-time multiplexing. In a

network printer, different users get their turn to use the printer one after another based on time.

Space Multiplexing

When a particular resource is space multiplexed, the resource is shared equally in the available

space. Time period for which a particular resource is used does not matter anymore. That is

each one gets a part of the resource. Hard disk, main memory, etc are some of the examples of

space multiplexing. These resources are most of the times divided equally between all the users.

To summarize, operating system is one of major programs, without which it would be almost

impossible for us to access the computer as we are accessing it these days. Without operating

system, computer would be nothing but a monitor executing binary and machine language data.

Glossary