Operating System Flashcards
Operating system
a program or set of programs that manage the operations of the computer for the user. It acts as a bridge between the user and the computer’s hardware, since a user can’t communicate directly with hardware. It disguises the complexities of managing and communicating with hardware from the user via a simple interface where they can click and immediately achieve tasks
where is OS held
permanent storage e.g. hard disk
the loader*
a small program held in ROM. When computer is turned on, the loader in ROM sends instructions to load the OS by copying it from storage to RAM
functions of OS
all OS share basic functions:
- processor scheduling
- UI
- provide utility software to carry out household tasks
- memory management and file handling
- ISR
- backing store management
- management of all I/O
- security for user files
how memory management works
each program/open file/copied clipboard item must be allocated a SPECIFIC area of memory whilst computer is running. When switching applications, both in memory simultaneously. If program closed, relinquishes memory and then free space. you don’t need to load the whole program into memory initally
paging
partitioning memory: memory divided into fixed page sizes eg 4kB and are physical divisions. a process held in memory may be held in several non-contigous pages. consecutive logical memory addresses physically stored in separate pages anywhere within physical memory and age table uses mapping to store link between physical memory address and virtual memory address.
used for virtual memory
page table does what
uses mapping to store the link between physical and logical memory address, preventing running out of memory and putting two things in the same place
why segmentation and paging
shuffling things in memory takes time so split up and makes better use of memory
both use VM and allow programs bigger than memory to run and allow non contiguous torage
segmentation
a way of partitioning memory by the logical division of memory into varying length segments of complete sections of programs depending on structure of program
virtual memory
Allows programs to run when insufficient memory available
as more jobs are loaded into memory OS may store PAGES (fixed sizes) of temoporarily inactive jobs onto disk as memory isn’t limitless. acts as secondary backing storage, an extension of memory to make room for pages next job with share of PROCESSOR TIME
large number of jobs loaded leads to
deterioration of performance as insuff. memory in RAM, pages swapped in and out until more time spent swapping than processing instructions- slows down PC (thrashing)
interrupts
a signal from a software program, hardware device or internal clock to the CPU as a request for processing time when requiring immediate attention, causing a break in the execution of the current routine. It allows important tasks to take precedence/be processed. Also allows multitasking
software interrupt
when an application program terminates or requests certain services from OS e.g. I/O operation complete, or error occurred e.g. printer out of paper
timer interrupt
indicates its time for next process to have processor time
memory management and why its necessary
To do with the organisation of main memory via converting logical addresses to physical addresses. OS manages allocation and management of space. Partitioning is used, involving paging/segmentation and virtual memory.
- allows memory to be Managed more easily
- allows Optimum use of memory
- allows programs to Share/be allocated memory and protects programs/data from each other
- allows you to work on Several at a time.
- allows running when Insufficient space in memory/allows running programs larger than main memory
first come first serve is inefficient
once a job starts- prevents other jobs being processed.
a job using slow resource e.g. printer wastes processor time
when a CPU receives an interrupt signal
It suspends execution of running program/process and lower priority interrupts. It puts values of program counter and each register and PC onto the system stack. Meanwhile source of interrupt is identified and ISR called to deal with interrupt- a particular routine will be run to service it depending on type of interrupt. after interrupt serviced original values of registers retrieved, interrupts are enabled and process resumes from point left off from
hardware interrupt eg
hard disk READ SET OF DATA BLOCKS
buffer is full or has free space
interrupts are assigned…
priorities (order of importance) and lower priorities may be disabled whilst a higher priority is being serviced, allows processes to take place and precedence
order of priority
power fail, clock, IO device sends signal requesting service or signalling end of IO operation
when are interrupt tests done
at end of each FDE cycle
processor scheduling
OS allocates processor time for each application as they compete for CPU because computers are able to run multiple applications simultaneously
a single processor can…
only process one instruction at a time but can multitask thanks to interrupts
single user multitasking
can run on a STANDALONE computer. only one user uses machine AT A TIME. constantly switching between small parts of multiple larger tasks in turn, according to SCHEDULING ALGORITHMS allowing more than one task (apparently) to be carried out at a time. giving appearance of carrying out several tasks simultaneously. Several tasks are in memory at one time but most not currently being executed
what is the scheduler*
the OS module responsible for ensuring processor time is used as efficiently as possible, allocating processor time to applications competing for CPU acc to scheduling algorithms
objectives of scheduler
- allow the processing, prioritisation of tasks and multitasking
- be fair to all jobs
- able to prioritise jobs and ensure they get processed (no starvation)
- maximum throughput (process as many jobs possible in least possible time)
- be fair to all users
- ACCEPTABLE RESPONSE TIME to all users (main aim)- maximise number of interactive users with fast/real time response time
- make max/efficient use of CPU time and other resources
- keep hardware as busy as possible
round robin*
> > PREEMPTIVE <>FIFO«_space;basis , and each have a limited/quantum very small time slice of CPU time, managed by interrupting timer generating interrupts.
If process not completed/higher priority occurs dispatcher gives CPU to next user/job and puts VALUES onto system STACK. Job put at back of queue.
Some systems have a system of priorities so higher priority has more than 1 consecutive time slice (order may depend on users priorities) Users are unaware of any delays
adv and disadv round robin
starvation free, all get a time slice
v
takes longer for all jobs
not always able to PRIORITISE
first come first served
jobs processed in order of arriving with NO SYSTEM OF PRIORITIES
disadv first come first served
if one large job comes all little ones have to wait a long time
shortest remaining time
a preemptive algorithm meaning process will be blocked by scheduler after a fixed time. Smallest estimated time to completion is run next
adv shortest remaining time and shortest job first
greater throughput/yield, less WAITING time, less smaller jobs BEHIND BIG JOBS, smaller jobs done quicker and advantages to new jobs
disadv of shortest remaining time and shortest job first
requires knowledge of how long a job will take so user has to estimate job time
large jobs may never get time share if smaller jobs keep coming (process starvation) or high priority interrupts
good jobs for shortest remaining time
batch jobs (group of records processed as a single unit without user inputs) eg payroll done regularly and overnight, scientific and commercial jobs run regularly
shortest job first
non-preemptive, won’t be blocked by scheduler after a fixed time. Shortest estimated time run next
multi-level feedback queue*
several job queues are implemented which jobs can move between: depending on how much processor time they use, characteristics of CPU bursts and need for CPU.
Give PREFERENCE to short jobs, IO bound processess
multilevel feedback queue: IO devices
IO devices are much slower than CPU so keep as busy as possible to prevent BOTTLENECK because SEVERAL PROGRAMS NEED to send data to same IO device. When 1 job using IO, another uses processor TO MAXIMISE PROCESSOR USE
how multilevel feedback queue works
algorithm chooses to process highest priority from corresponding queue and runs. If too much time, moves to lower priority and if waiting too long, moves higher queue
adv multilevel feedback queue*
SHORT CPU jobs prioritized and processed quickly, IO bound processes given preference. maximises CPU use, no starvation, preemptive and non preemptive
disadv multilevel feedback queue
requires MONITORING and COSTLY
backing store management*
OS needs to keep directory of where files are stored and which areas of storage are free so files can be accessed quickly and knows where files and applications can be stored. When files/applic are loaded- transferred FROM backing storage into memory. uses a file management system
file management system
COMES WITH desktop OS enabling users to move/delete and protect files and folders from unauthorised access. disguises complexities of managing backing storage
what is backing storage
name of all other storage devices aka secondary storage which is non-volatile that is used to store data for a long time
main memory
for storing data temporarily which is volatile and stores data and programs whilst in use and being PROCESSED by CPU
peripheral management*
different APPLICATIONS require different I/O devices throughout operation. OS communicates with IO to ensure switched on, online, correct device and send correct data
when data is sent to IO device
transferred to buffer (area of memory temporarily holding data, that compensates for the difference in speed between IO device and processor) so CPU can continue with another task ALLOWING MAX USE OF CPU
interrupts in peripheral management
to know when buffer is filled/emptied so it is requesting for more data to be transferred or to stop sending data
distributed OS*
a form of parallel processing that splits and spreads the workload/processing load over multiple, independently networked, physically separate computer nodes/servers/ PROCESSORS. Tasks are run SIMULTANEOUSLY and results are then gathered together, so it APPEARS to be a single system to the user
distributed OS is controlled by
a system management software that brings all INTERCONNECTED systems together and coordinates individual and collaborative activity on the different nodes
intranets
system configured as a cluster of servers that share MEMORY AND TASKS, providing more POWER than a single large server = better PERFORMANCE
mainframe
computers with lots of processors and huge storage simultaneously serving thousands of users (expensive)
supercomputers
mainframe designed to go fast than store a lot eg gaming
disadv parallel processing
it has to be determined what can actually be split
multiuser OS
a single powerful mainframe/supercomputer connected to dozens or hundreds of terminals all using mainframe CPU simultaneously, sharing resources. Each user gets a slice of processor time according to an ALGORITHM
multiuser OS manages
each user logged onto system and workspace
allocates resources to jobs wanting to run
keeps logs of resources and processing time jobs use
working out most efficient use of processor CYCLES
maintaining SECURITY
mobile phone is a
multitasking computer with its own OS
mobile phone OS
a OS used on phones/tablets/PDAs and other mobile devices usually TIED TO SPECIFIC HARDWARE combining features of a PC OS with own special features useful for mobile use eg cellular. Made of main system and low level propriety real time OS
mobile phone OS special features
responding to users touch, speech recognition, camera, GPS, music player
main system mobile OS
operates UI and application software
real time mobile OS
radio and other hardware but range of security vulnerabilities permitting others to gain control over mobile device
embedded OS
have a dedicated function and usually part of a larger device. Interfaced in a more bespoke manner. Low POWER consumption/COST and use limited HARDWARE and RESOURCES
eg embedded OS
washing machine, microwave oven (app program in ROM) control system of passenger aircraft
embedded OS interface in household appliance
minimal UI
accepts inputs from sensors and sends outputs to control devices
no or little RAM and no memory management system
no permanent data storage to be managed
why are priorities necessary for interrupts
order of importance- to ensure most EFFICIENT use of processor
to ensure most urgent task performed first
to allow tasks to take place and precedence
to choose which interrupt to process if more than one occur together
to decide between interrupt and task
priorities can change because
allows it to be processed otherwise if higher priorities keep occurring, it would wait indefinitely
transfer of memory between main memory and hard disk drive
buffer filled from primary memory and processor can continue with other tasks. Buffer emptied to hard drive. Interrupt sent to processor to request buffer refilled. Interrupt priority determines position in interrupt queue. Process continues until all data transferred.
network OS facilities
file handling and save in a remote location
communications between machines
sharing resources eg hardware, software
real time OS
OS running under pre-defined deadlines, completing tasks in a guaranteed time. for tasks classed as time critical where right answer too late is wrong
eg RT OS
aircrafts
RT OS features
- able to respond quickly to inputs and sensors
- must be able to deal with many inputs simultaneously
- must have a failsafe mechanism to detect and take appropriate action if hardware component fails
- must incorporate redundancy: if a component fails: automatically switches to backup hardware
BIOS
basic input output system: firmware designed to run when PC is first switched on. Its responsible for detecting, initializing and testing hardware components and booting OS. historically provided ABSTRACTION layer allowing a consistent way for application programs and OS to interact with IO devices
how does BIOS boot OS
copies (key parts) of OS from hard disk to RAM
where is BIOS stored
EPROM- erasable programable ROM
device drivers*
a computer program providing a software interface to a particular hardware device enabling OS to access hardware functions without needing to know details of the hardware HARDWARE DEPENDENT AND OS SPECIFIC
device drivers installation and communication
they may have to be isntalled but OS often comes with common drivers. they are HARDWARE DEPENDENT AND OS SPECIFIC and communicate with devices via system bus/communications subsytem
how device drivers work
calling program invokes a routine in the driver,
driver issues command to device, device sends data back to driver, driver potentially invokes routines in original calling program
virtual machine
any instance where software is used to take on the function of a machine including executing intermediate code or running an OS within another
VM user experience
UE is the same on a VM as a dedicated machine. but first choose OS to use
benefits of VM
saves costs as reduces need for separate hardware and provides more SECURITY when executing intermediate code
virtual device drivers
fools OS into thinking its directly communicating with a certain piece of hardware when actually going through function calls to host
