Operating Systems: Introduction

Introduction

• Chapter 1 introduces the fundamental concepts of Operating Systems (OS).
• It provides an overview of the structure, operations, and goals of an OS.
• The chapter outlines key aspects such as computer system organization, memory management, and security.

Objectives

• The basics of computer system organization are described.
• A summary of the major components to OSs is provided.
• An overview of different computing environments is included.
• Open source OS are explored.

What is an Operating System?

• An OS serves as an intermediary between computer users and hardware.
• OS objectives include executing user programs, simplifying problem-solving, ensuring convenient use, and using hardware efficiently.

Computer System Structure

• The structure can be divided into hardware, operating system, application programs, and users.
• Hardware provides computing resources (CPU, memory, I/O devices).
• The OS controls and coordinates hardware use across applications and users.
• Application programs utilize system resources to solve user computing problems including word processors, compilers, web browsers, database systems and games.
• Users can be people, machines or other computers.

What Operating Systems Do

• The importance of convenience, ease of use, and good performance for users are important for computer systems.
• Shared computers like mainframes must keep its users happy, therefore resource utilization is important
• Handheld computers are resource poor, therefore usability and battery life are important.
• Embedded computers in devices and automobiles may have very little to no user interface.

Operating System Definition

• An OS functions as a resource allocator and manages all hardware resources.
• It decides between conflicting requests for efficient and fair resource use.
• The OS serves as a control program that controls program execution to prevent errors.

Operating System Definition (Cont.)

• There isn't one accepted definition for an OS.
• A good approximation is "Everything a vendor ships when you order an operating system."
• The one program running all the time is the kernel.
• Everything else is a system program, which ships with the OS, or an application program.

Computer Startup

• The bootstrap program boots the system at power-up or reboot.
• Commonly stored in ROM or EPROM, also known as firmware.
• It initializes the system and loads the OS kernel to start execution.

Computer System Organization

• Computer system operations include one or more CPUs with device controllers who connect through a common bus and provide access to shared memory.
• CPUs and devices can have concurrent execution, but may compete for memory cycles.

Computer-System Operation

• I/O devices and the CPU can execute concurrently.
• The device controller is responsible for a specific device type and has a local buffer.
• The CPU moves data to and from main memory and local buffers.
• The I/O goes from the device to the local buffer of the controller.
• The device controller informs CPU when its operation is over by sending an interrupt.

Common Functions of Interrupts

• Interrupts transfer control to an interrupt service routine, typically via an interrupt vector containing addresses of service routines.
• Interrupt architecture must save the address of the interrupted instruction and return there later.
• A trap or exception is a software-generated interrupt caused by either an error or user request.
• An operating system is interrupt driven.

Example: Reading from disk

• CPU fetches a read instruction from memory into the instruction register and decodes it.
• If data are needed, the CPU brings them from memory to data registers.
• The program counter holds address of what will be executed next and then the CPU begins executing the read instruction.
• A system call is made when code calls the read command which requests usage of HD.
• This is considered an interrupt because the CPU is requesting a system service and stops its current program temporarily.
• The source of the interrupt, because it comes from the program/software, is considered a trap or exception.
• The HD driver then sends a signal via the bus to request data from the HD.
• The read request signal is provided to the disk controller and the data is read and transferred to the HD buffer.
• The HD controller sends an interrupt signal via bus to let the CPU know the operation is done.
• A software interrupt is performed by a system call, which enables a switch from user mode to system mode.

Interrupt Handling

• The OS saves the CPU state by storing registers and the program counter to handle interrupts.
• The type of interrupt is determined through polling, where the CPU checks the device status, or through a vectored interrupt system, where the device controller informs the CPU.
• Separate code segments define actions for each interrupt type.

Interrupt Timeline for a Single Process output

• CPU runs process 1, and during this, the I/O device is idle.
• Process 1 starts an I/O, so the device starts working and the CPU switches to another process.
• The I/O device is then idle after it finishes, issues an interrupt for the CPU,
• Then the CPU stops working on the other job, takes the interrupt call, and goes back to the first process.
• This repeats until the process 1 finishes everything.

I/O Structure

• After I/O starts, control returns to user program after I/O completion.
• The CPU idles until the next interrupt.
• For system calls, the OS indexes into the table to see the type and status to modify it.

Storage Definitions and Notation Review

• The basic unit of computer storage is a bit (0 or 1).
• A byte is 8 bits
• Kilobyte (KB): 1,024 bytes
• Megabyte (MB): 1,024^2 bytes
• Gigabyte (GB): 1,024^3 bytes
• Terabyte (TB): 1,024^4 bytes
• Petabyte (PB): 1,024^5 bytes
• Computer makers often round numbers; 1 million bytes (MB), and 1 billion Bytes (GB)

Storage Structure

• Main memory is the only large storage media that can be directly accessed by the CPU.
• Is random access memory (RAM) and volatile (lose data when power is off).
• Secondary storage is an extension of the main memory which provides large nonvolatile storage.
• Hard disks are rigid metal or glass platters with magnetic recording material, and are logically divided into tracks and sectors.
• The disk controller, determines the interaction of the device and computer.
• Solid state disks are faster than hard disks and non-volatile.

Storage Hierarchy

• Storage is organized in a hierarchy of speed, cost, and volatility.
• Caching copies data into faster storage.
• Main memory can be used as a cache for secondary storage.
• The device driver manages I/O and makes the interface between the controller and kernel uniform.

Caching

• Caching is a way to copy quicker information into a faster storage location temporarily.
• The cache checks the faster storage to see if the the information is there.
• If it is there, the cache will use information right from the cache.
• If it is no there, it will copy the data over to cache and then use the information.
• An important design choice behind this, is that cache is much smaller than the storage it is being cached.
• There are important design problems that are considered with this, cache size and replacement policy, and cache consistency.

Direct Memory Access (DMA) Structure

• Used for high-speed I/O devices that can transmit data close to memory speeds.
• The buffer storage's blocks of data are transferred directly to main memory by the device's controller.
• Instead of having an interrupt for every byte, and DMA uses only one interrupt per block.

How a Modern Computer Works

• Fetch instructions from memory, place them in the Instruction Register, decode them, and execute them.
  • This is part of Von Neumann architecture.

Computer-System Architecture

• Multiprocessor systems are on the rise.
• Multiprocessors are known as tightly-coupled or parallel systems.
• Advantages include increased throughout, more economy, and increased reliability.
• There are two types of multiprocessing
  • Asymmetric which gives each processor a specific task, and Symmetric which enables each processor to conduct all tasks.

A Dual-Core Design

• A chip can have one or multiple cores.
• Multichip can have several chips with one core.
• Multicore is when you have several cores on one chip.
• Communication of multi-cores on a single chip is faster than multi-chip communication.
• Some systems can contain all chips within a computer.

Clustered Systems

• It is like multiprocessor systems, but with multiple systems that work together by sharing local area networks.
• They offer high-availability through asymmetric or symmetric clustering.
• High-performance computing (HPC) relies on clusters.
• A Storage Area Network (SAN) is a network of switches, routers, and storage devices that are used to provide connectivity for hosts.

Operating System Structure

• Multiprogramming is a batch system that is used for efficiency to execute multiple programs simultaneously.
• Multiprogramming organizes jobs (code and data), therefore the CPU always has something to execute
• Time sharing provides a logical extension in which the CPU switches jobs frequently, so users can simultaneously interact with their jobs in interactive computing.
• If there are several jobs that are ready to run at the same time, CPU scheduling kicks in, if there is not enough memory, then swapping moves a process, and then virtual memory allows execution of processes not completely in memory.

OS Features Needed for Multiprogramming

• These features include supplied I/O routines, allocating jobs memory management, disk job scheduling, and in-memory CPU scheduling.
• In addition, there needs to be allocation of devices and protection of both the OS and the programs within.

Operating-System Operations

• Operations in the OS are interrupt driven by both hardware and software.
• Software interrupts include software error, a requested OS service, and process issues such as modifying each other, or infinite loops.

Operating-System Operations

• Dual-operation modes allow the OS to protect itself with user mode and kernel mode enabled via a mode bit within hardware.
• If in kernel mode, this gives special instruction to conduct tasks, and conducts all I/O instructions which then changes to kernel, and reverts back to user when complete.
• VMs are supported with multiple modes for guest Machines.

Transition from User to Kernel Mode

• Timers set to prevent infinite loops or a process hogging resources, where after, counter zero generates an interrupt.

Common Operating-System Components

• The operating system has common items: process, memory, and storage management, file and I/O subsystem implementation, and security/protection tools.

Process Management

• Active entities, processes take the form of programs.
• They need CPU, memory, and I/O resources to accomplish their task
• After it is over, termination reclaims re-usable resources.
• Process activities include creating and deleting processes, suspending and resuming processes, mechanisms for process synchronization, communication, and deadlock.
• There are two types of processes, processes with a single thread, or multiple threads.

Memory Management

• Every program requires all or part of their instructions to be in-memory for execution.
• What data is needed and when is determined by memory management and helps overall response time.
• Activitiees include keeping track of memory usage, which processes/data to move, and space allocation which optimizes CPU utilization and response time.

Storage Management

• The OS has a physical to logical storage mapping in the file.
• The OS is device controlled with access based on speed, capacity and data-transfer rates.
• Storage management occurs through file system management.
• They are put into directories.
• Access is controlled by user/group levels, and primitives exist to manipulate them.
• Stable storage is created through backups.

Mass-Storage Management

• Disks store necessary memory not stored elsewhere.
• Proper management is critical, therefore OS activities include managing disk scheduling, allocation, and freespace.
• Tertiary storage and WORM storage also have to be managed.

Performance of Various Levels of Storage

• Fast storage=lower access time, means high bandwidth.
• Registers are small, have the lowest access time, magnetic disks the highest.

Migration of data “A” from Disk to Register

• Because in a multitasking environment the most recent value is critical, cache coherency is critical and has to be managed within the software.

I/O Subsystem

• The OS hides the specific functionality of hardware devices from the user, and manages the I/O, which includes.
• Buffering.
• Caching.
• Spooling.

Protection and Security

• Protection ensures processes or users have authorized access.
• Security defends various attacks to the system.
• Systems distinguish which users can use what and protect against illegitimate access.
• Protection is done through user IDs + appropriate access control to processes and files.
• Group identifiers also exist.

Kernel Data Structures

• Linked lists, double linked lists, circular linked lists, and Binary search trees are common.
• Binary search trees on the right are greater and search can be (n), also balanced search trees.
• Bitmap string is a 1 or 0 string representing the status of the ite