Functions of Operating Systems
What the Specification Says
Describe the main features of operating systems, for example memory management, and scheduling algorithms;
Explain how interrupts are used to obtain processor time and how processing of interrupted jobs may later be resumed,
(typical sources of interrupts should be identified and any algorithms and data structures should be described);
Define and explain the purpose of scheduling, job queues, priorities and how they are used to manage job throughput;
Explain how memory is managed in a typical modern computer system (virtual memory, paging and segmentation should be described along with some of the
problems which could occur, such as disk threshing);
Describe spooling, explaining why it is used;
Describe the main components of a typical desktop PC operating system, including the file allocation table (FAT) and how it is used, and the purpose of the boot file.
Notes
Operating systems must:
§ Provide and manage hardware resources
§ Provide an interface between the user and the machine
§ Provide an interface between application software and the machine
§ Provide security for data on the system
§ Provide utility software to allow maintenance to be done
Why OS’s were developed:
§ If a program needed an input or an output the program would have to contain the code to do this. There were often multiple occurrences.
§ Short sub-routines were developed to carry out these basic tasks, like get the key pressed.
§ The joining together of these sub-routines lead to the input-output-control-system (IOCS). As this started to get more complex it turned into the OS.
§ Also as programs started to be written in higher-level languages translators were needed, these were incorporated into the OS
Types of Interrupts
§ I/O interrupts – generated by the I/O devices to signal a job is complete or an error has occurred
§ Timer interrupt – generated by internal clock to signal that the processor should attend a time critical activity
§ Hardware error – for example power failure, where the computer must try to shut down as safely as possible.
§ Program interrupt – generated due to an error in a program, such as violation of memory usage.
Dealing with interrupts
§ Interrupts are stored in queues
§ The position in the queue is determined by its importance to the user, system and job – priority.
Objectives of Scheduling
§ Maximise the use of the whole computer system
§ Be fair to all users
§ Provide reasonable response time to all users – weather online or batch processing users
§ Prevent the system failing or becoming overloaded
§ Ensure the system is consistent by giving similar response times to similar
Deciding which order to schedule jobs:
§ Priority – where some jobs have higher priority than others
§ I/O bound jobs – if a job is bound to a device, and is not served first it may prevent the devise being served efficiently.
§ Type of job – batch processing and real time jobs require different response times
§ Resources used so far
§ Waiting time – the amount of time a job has been waiting to use the processor
Ready, running and blocked queue
§ A job may be in any one of three states
§ Ready
§ Running
§ Blocked – e.g. waiting for a peripheral devise
What is a scheduler?
The scheduler is just a program, a set of instructions, used by the OS to decide where each of the jobs which are in its control should be and in what order to manipulate them.
Types of schedulers
§ High-level scheduler (HLS) – places job in the ready queue, makes sure that the system is not overloaded.
§ Medium-level scheduler (MLS) – swaps jobs between the main memory and backing store.
§ Low-level scheduler (LLS) – moves jobs in and out of the ready state, and decides which order to place them in the running state
Pre-emptive and non-pre-emptive schedulers
§ Non-pre-emptive - puts a job in the running state then leaves it there until the job does not need to run any more but needs to go to one of the other states, for example it may wait for an input or may have finished executing altogether
§ Pre-emptive - has the power to decide that the job that is running should be made to stop to make way for another job
Methods of deciding what order jobs should go in
§ First com first serve (FCFS) – the first job to enter the running queue is the first job to enter the running state. Favours long jobs.
§ Shortest job first (SJF) – sort the jobs in the ready queue in ascending order of time needed
§ Round robin (RR) – gives each job a maximum length of processor time, it will then go to the back of the queue again. Once its finished, it leaves the queue.
§ Shortest remaining time (SRT) – the ready queue is sorted on the amount of expecting time still to do. Favours short jobs, and there’s a danger long jobs may never be started.
§ Multi-level feedback queues (MFQ) – involves several queues of different priorities with jobs migrating across
Memory management
In order for a job to be processed, it must be stored in the memory – obviously. If several job are stored in the memory, their data must be protected from the actions of other jobs.
How free memory can be allocated
If a small gap in the memory becomes free, but it is not big enough for the next job, there are a number of options:
§ All the jobs could be moved upwards, so that all the bits of free space are together, and the next job can fit. Although this would use a lot of processing power, as all the addresses would have to be recalculated.
§ Another solution would be to split up the new job into available free memory. Although doing this could cause the jobs to all get split up into many tiny parts.
Linkers and Loaders
§ The loader – small utility program that loads jobs and adjusts addresses.
§ The linker – links parts of the program together
Linkers and loaders are integral to the management of memory space
Paging and Segmentation
§ Pages are equal sized sections, where a job will be allocated a number of pages to store itself in. They may be in order, or out of order.
§ Segments are of variable size to match the size of the job
Virtual Memory
When data is not needed to be accessed immediately, or programs have been ‘minimised’ for a long time, they may be transferred to virtual memory so as not to fill up the main memory. Virtual memory is on the hard disk, although it behaves exactly like the RAM.
Disk Thrashing
This is where the code contains many jump instructions so the processor spends most the time switching between the virtual memory and main memory. It involves the disk continuously searching for pages.
Spooling
This is where an I/O devise is a lot slower than the processor, so the job is moved into another storage location while it is being used. It is kept track of in a spool queue. A spool queue is only a reference to the jobs, and not a proper queue. A spool queue also allows for priorities, more important jobs can push into the queue and be done first.
Multi-tasking OS
There are two main types of OS, command driven e.g. MS DOS, and GUI. All OS’s allow the user to do basic things like rename, delete and move files stored in a hierarchical structure on the disk. Most GUI’s also allow for much more than this, they let the user multi-task, where several programs are running.
File Allocation Table (FAT)
§ This is a table that uses a linked to point to the blocks on the disk that contain files.
§ In order for this to be done, the disk must first be formatted, which involves dividing the disk, radially, into sectors and into concentric circles called tracks. Two or more sectors on a single track make up a cluster.
§ To find a file, the OS looks for the file name, and gets the corresponding cluster number, which can be used to find the file.
§ When a file is deleted, the clusters that were used to find the file can be set to zero.
§ The FAT table is usually loaded into the RAM to speed things up.
Loading and OS
§ PC is switched on
§ Contains very few instructions
§ Runs the power on self-test (POST) which resides in the permanent memory
§ The POST clears the registers, loads the address of the first instruction in the boot program into the program counter
§ The boot program is stored in the read-only memory and contains the skeleton of basic input output system. The BIOS structure is stored in ROM.
§ The user-defined parts of the BIOS are stored in the CMOSRAM.
§ The CPU then sends signals to check that all the hardware is working correctly
§ This includes checking busses, system clock, RAM, disk drives and keyboard.
§ If any of these devises contain their own BIOS then this is incorporated into the system’s BIOS
§ The PC is now ready to load the OS
§ The boot program checks if a disk is present for drive A, if so it looks for an OS on that disk. If no OS is found, an error message is produced.
§ If there was no disk in drive A, the boot program will check drive C.
§ Once found the OS, if it is Windows, the boot program will look for MSDOS.SYS, and IO.SYS which holds the extensions to the ROM BIOS and contains a routine called SYSINIT which controls the rest of the boot procedure.
§ SYSINIT then takes control and loads MSDOS.SYS which works with the BIOS to manage files and execute programs
§ The OS searches the root directory for a boot file such as CONFIG.SYS which tells the OS how many files may be opened at the same time and instructions on how to load the necessary devise drivers.
§ The OS tells MSDOS.SYS to load a file called COMMAND.COM, which is in three parts. The first part is a further I/O extension which joins with the BIOS to become part of the OS. The second part contains resident OS commands such as DIR and COPY.
§ The files CONFIG.SYS and AUTOEXEC.BAT are created by the user so that the PC starts up in the same configuration each time it is switched on.
Keywords and Definitions
§ Operating System (OS) – A piece of software that provides a platform on which the applications software can run. It controls the hardware, and provides communication with the outside world.
§ Interrupt - messages sent to the processor by some external entity asking them to stop what they are doing and do another job
§ Scheduler - a program, or set of instructions, used by the OS to decide where each of the jobs which are in its control should be and in what order to manipulate them
§ High-Level Scheduler (HLS) – scheduler that places the jobs in the ready queue, and ensures that the system never gets overloaded.
§ Medium-Level Scheduler (MLS) – Scheduler that swaps jobs between main memory and backing store.
§ Low-Level Scheduler (LLS) - moves jobs in and out of the ready state, and decides which order to place them in the running state.
§ Pre-Emptive Scheduler – scheduler that is allowed to move jobs out of the running state and into the ready queue. Has the power to move jobs.
§ Non-Pre-Emptive Scheduler – scheduler that does not have the power to move jobs until they are finished.
§ First com first serve (FCFS) – a priority method where the first job to enter the running queue is the first job to enter the running state. Favours long jobs.
§ Shortest job first (SJF) – a priority method where jobs are sorted in the ready queue in ascending order of time needed
§ Round robin (RR) – a priority method which gives each job a maximum length of processor time, it will then go to the back of the queue again. Once its finished, it leaves the queue.
§ Shortest remaining time (SRT) – a priority method where the ready queue is sorted on the amount of expecting time still to do. Favours short jobs, and there’s a danger long jobs may never be started.
§ Multi-level feedback queues (MFQ) – a priority method which involves several queues of different priorities with jobs migrating across
§ Memory Management - One of the key jobs of the OS, managing memory and allocating it accordingly to jobs.
§ The Linker – a utility program included in the OS which links parts of the program together through the memory.
§ The Loader – a utility program which loads jobs into the memory and adjusts addresses accordingly.
§ Pages – equal sized sections in the memory, which are fixed and jobs are allocated a number of pages.
§ Segments – similar to pages, although the size can be variable to match the job
§ Virtual Memory - is part of the hard disk, but has faster speeds, it is slower than the RAM, but is usually considerably bigger. Acts in the same way as RAM
§ Disk Trashing – caused by many jump instructions in the code, requiring main memory and virtual memory to be repeatedly switching. The hard disk will have to continuously search for pages.
§ Spooling -It is a method used to place input and output on a fast access storage device, such as a disk, so that slow peripheral devices do not hold up the processor. It allows for queues when several jobs want to use peripheral devices at the same time. It stops different input and outputs becoming mixed up.
§ Command Driven OS – an operating system without a GUI for example MSDOS
§ GUI OS – An OS with a GUI, like most modern ones e.g. Windows 8
§ Multi-tasking OS – most GUI OS’s support multi-tasking, where multiple jobs can be carried out simultaneously.
§ File Allocation Table (FAT) - a table that uses a linked lists to point to the blocks on the disk that contain files, the actual table is usually stored in the RAM.
§ Power on self-test (POST) – a routine the resides in the permanent memory, it clears the registers, loads the address of the instruction into the program register. It then sends signals to the necessary hardware to check it’s all there and working properly.
§ Basic Input Output System (BIOS) - instructs the computer on how to perform a number of basic functions such as booting and keyboard control.
§ CMOS RAM– stores the user-defined part of the BIOS. The BIOS is customisable.
§ Boot Program – gets the system ready to accept an operating system. Unalterable and stored in the ROM.
§ Boot File - contains some basic parameters to which the system will operate. It can be altered, and is stored in the CMOS RAM.
§
Key Points Posters
Spooling
Loading an OS
Past Exam Questions with Mark Scheme Answers
Why interrupts are used in a computer
system
|
to obtain processor
time... |
State some sources of
interrupts
|
(imminent) power failure/system
failure |
Why is scheduling used?
|
maximise number of
users… |
Why are jobs given different priorities in a job
queue?
|
some jobs are more urgent than
others |
Why is memory management
necessary?
|
to allocate memory to allow separate
processes to run at the same time |
Why is virtual memory
needed?
|
to allow programs to run that need more
memory than is |
How is virtual memory used?
|
use of backing store as if it were
main |
What's the problem of disk
threshing?
|
occurs when using virtual memory/moving
pages |
Methods of scheduling
|
round robin |
What is spooling and when is it
used?
|
output data to disk drive/storage
device |
How are paging and segmentation
similar?
|
ways of partitioning memory
|
How are paging and segmentation
different?
|
segments are different sizes but pages
are fixed size |
What
problems may occur from paging and segmentation?
|
disk threshing |
When is the boot file used?
|
while the operating system is loading
|
What is the purpose of the boot
file
|
provides personal settings |
What’s virtual memory
|
use of backing store…
|
What’s the purpose and use of the file
allocation table?
|
a map of where files are
stored… |
Further Recourses
VRS(http://vrs.as93.net)
§ Prezzi Presentation
§ More revision posters
§ More past exam questions
§ Links for further reading
§ Key words list
§ Revision Quizzes
§ This document
§ User contributed content
Revision Quizzes (http://RevisionQuizzes.com)
§ Jobs the OS must do quiz
§ Loading an OS quiz
§ Functions of the OS summary quiz
Links