What is an OS

Reference: Ewha Womans University OS Lecture - Professor Ban Hyo-kyung

What is an Operating System?

• It provides a convenient environment for users to use the computer system without knowing the hardware

• It efficiently manages the limited resources (CPU, Memory, I/O) of the computer system to achieve maximum performance

  • ex)

    • Allocating CPU to running programs in short time intervals alternately (round robin)

    • Appropriately distributing memory space to running programs

      • memory: the workspace of the CPU

• A software layer installed directly on top of computer hardware that connects users and all other software to the hardware

  • Narrow definition (Kernel)

    • The core part of the operating system that resides in memory

  • Broad definition

    • A concept that includes not only the kernel but also various peripheral system utilities

Functions of the Operating System

• The computer boots up

  == the operating system is loaded into memory

• Kernel

  • Resides in memory when the computer is running

• CPU Management

  • Reads machine instructions from memory every clock cycle and performs computations

  • The CPU cannot directly access I/O devices

    • Each I/O device has its own small CPU-like I/O controller, and the CPU requests the controller to execute the necessary operations on the I/O device

  • CPU Scheduling

    • The most important function of the operating system!

    • Machine instructions within the operating system command the CPU to execute specific programs, i.e., scheduling

    • A specific program is allowed to use the CPU for a limited time only, and when the allotted time expires, the CPU is handed over to another program

      → The operating system cannot do this alone; it requires hardware assistance

• Memory Management

  • How to divide and use limited memory

• Disk Scheduling

  • Disks are slow!

  • While executing a long-running task (e.g., reading a file from disk), the CPU moves on to another task and executes it

  • Just like the CPU, multiple requests come into I/O devices simultaneously, and the operating system schedules the order in which to process them efficiently so resources can be used effectively

• Interrupt and Caching

  • Interrupts and Caching are used to buffer the speed difference between the fast CPU and slow I/O devices

  • Caching

    • Introducing an intermediate layer

    • Caching frequently accessed data in memory

  • Interrupt

    • The CPU assigns work to the I/O controller and does not wait; instead, it finds a task that can be executed immediately and runs it

      • The I/O controller notifies the CPU that the task is complete through an Interrupt

        • The CPU performs an Interrupt check after each task to verify whether executed tasks have finished

    • When an Interrupt arrives, the CPU is handed over to the operating system, and the completed task is made available for the CPU to execute again

    • Interrupts are a mechanism for slow devices (e.g., disk I/O) to notify the CPU that their tasks have been completed

Purpose of the Operating System

• Provide a convenient environment for using the computer system

• Efficiently manage the resources of the computer system

  • The brain of the computer is not the CPU but the OS! (professor's opinion)

    • The CPU is an entity good at calculations

    • Memory is an entity that stores things

    • The OS is an entity that decides how to use them (the ruler)

      • e.g.) When memory needs to be appropriately divided among programs, the OS makes that decision

Classification of Operating Systems

Concurrent Task Capability

• Single tasking

  • Processes only one task at a time

• Multi-tasking

  • Processes two or more tasks simultaneously

  • e.g.) Another command or program can be executed before one command finishes

Number of Users

• Single user

  • e.g.) MS-DOS, MS Windows

• Multi-user

  • e.g.) Unix, Linux

  • Security and permission issues need to be resolved to support multiple users

Processing Method

• Batch processing

  • Collects a certain amount of work requests and processes them all at once

  • Must wait until the task is completely finished

  • e.g.) Early Punch Card processing systems

• Time sharing

  • When performing multiple tasks, the computer's processing power is divided into fixed time units

  • Has shorter response time compared to batch processing systems

  • An interactive method

  • A method of splitting time to use given resources as efficiently as possible

  • e.g.) Unix

• Realtime OS

  • An OS for real-time systems where certain tasks must be guaranteed to complete within a specified time

  • Extended concept of real-time systems

    • Hard realtime system

      • Missing a deadline causes serious problems

      • e.g.) Nuclear reactor / factory control, missile control, semiconductor equipment, robot control

    • Soft realtime system

      • Best effort → tries its best but cannot guarantee

      • e.g.) When playing a video, if frames are not rendered, the video stutters

Terminology

The following 4 terms are fundamentally the same in meaning, but each emphasizes slightly different aspects

• Multi-tasking

• Multi-programming

  • Emphasizes that multiple programs are loaded in memory simultaneously

• Time sharing

  • Emphasizes the meaning of dividing CPU time and sharing it

• Multi-process

The following term has a completely different meaning

• Multi-processor

  • Means that multiple CPUs (processors) are attached to a single computer

  • A topic dealt with in high-performance computing and cloud computing

Examples of Operating Systems

UNIX

• Most of the code is written in the C language

  • To control hardware elements, the operating system needs to be written in assembly language, but since that is difficult, the C language was created for easier human understanding

  • Became easier for humans to understand

• High portability

  • C language is compatible across multiple architectures, providing high portability

• Minimal Kernel structure

• Easily extensible for complex systems

• Source code is open

• Suitable for program development

• Various versions

  • System V, FreeBSD, SunOS, Solaris

  • Linux

    → Source code is open and development continues based on it

DOS (Disk Operating System)

• Developed by MS for IBM-PC in 1981

• Single-user operating system

• Limited memory management capability (main memory: 640KB)

MS Windows

• MS's multi-tasking GUI-based operating system

• Plug and Play

• Enhanced networking environment

• Provides compatibility with DOS application programs

• Rich support software

Structure of the Operating System

Process Management

• Process creation and deletion

• Resource allocation and deallocation

• Cooperation between processes

Others

• Protection system

• Networking

• Command line interpreter

Process

• What is a process?

  • A running program

• CPU Queue

  • Programs that want to use the CPU are lined up in a queue

CPU Scheduling

Determining the order in which tasks in the CPU queue are executed when multiple programs are waiting in the CPU queue to use the CPU

→ One of the important roles of the operating system

FCFS (First-Come First-Served)

• Description

  • A method of processing in the order requests arrive

    → Good fairness

• Problem

  • If an earlier request occupies the CPU for a long time, the average waiting time for subsequent tasks increases

    → Not efficient

SJF (Shortest-Job-First)

• Description

  • Schedules the process with the shortest CPU usage time first

  • Guarantees minimum average waiting time

    → Good efficiency

• Problem

  • Starvation may occur

    → Fairness issue

Round Robin (RR)

• The most widely used method in modern systems

• Each process has an equal CPU allocation time

  • When the allocation time expires, an Interrupt occurs, the process loses the CPU, and it goes to the back of the CPU queue

• When n processes are in the CPU queue

  • No process waits more than (n-1)*allocation time

    • Total processes n - myself

• Advantages

  • Short tasks can complete quickly within the given allocation time, and long tasks can continue their unfinished work when their turn comes again

    → Waiting time is proportional to the process's CPU usage time

Memory Management

• Swap area

  • The immediately needed space of a running process is loaded into memory, but the parts that cannot fit due to memory space limitations are placed in the swap area

  • Located on a persistent disk, but the data becomes useless when power is cut off

    • Because the corresponding process is no longer running

• What should be evicted from memory?

  • Pages likely to be used again in the future should not be evicted, and pages unlikely to be used in the near future should be evicted for efficient management!

• LRU (Least Recently Used)

  • Deletes the page that was referenced the longest time ago

• LFU (Least Frequently Used)

  • Deletes the least frequently referenced page

Disk Scheduling

• The disk head movement time takes the most time among disk access times

  → Scheduling should be done to reduce head movement distance

• Seek time must be reduced!

  → Disk scheduling is needed for this purpose

FCFS (First-Come First-Served)

Takes a long time (inefficient)

SSTF (Shortest Seek Time First)

• Efficient due to short movement distance

• but, starvation problem occurs

SCAN

• Current disk scheduling method

• The head moves regardless of requests in the queue

• Processes requests from the queue as it moves

  → Similar to how an elevator operates!

Storage Hierarchy and Caching

• The CPU executes machine instructions using values in Registers

• Cache Memory is placed between the CPU and memory to overcome the speed difference

• Currently, Flash memory is positioned between Main Memory and Magnetic Disk

• Primary

  • Volatile

  • CPU can directly access it (executable)

• Secondary

  • Non-volatile

  • CPU cannot directly access it, so it must be loaded into Primary for execution

• A layered architecture is used to overcome (buffer) speed differences

  • Copying data to a faster layer and reading the copy is exactly what Caching is!

• Caching

  • copying information into faster storage system

  • The question is: what to evict!

Flash Memory

What is Flash Memory?

• A semiconductor device (hard disk: magnetic)

• NAND type (storage), NOR type (for embedded code storage)

  • What we use (for storage purposes) is all NAND type!

Characteristics of Flash Memory

• Non-volatile

  • Data is preserved even when power is turned off

• Low power consumption

  • Consumes less power compared to hard disks

• Shock resistance

  • Resistant to physical impact

• Small size

• Lightweight

• Write count limitation

  • Cannot write anymore after a certain number of writes

• Data can degrade over time

  • Data is read based on the amount of charge in a cell, and charge leaks out of the cell over time

    • Eventually, the data may become unreadable

Usage Forms of Flash Memory

• For embedded system configuration in mobile phones, PDAs, etc.

• USB memory sticks

• SD cards

• Not only for mobile devices but also attempting to replace hard disks in large-scale systems under the name SSD (Solid State Drive)