🖲️Operating Systems Unit 5 – Input/Output Systems

What Are I/O Systems?

• Handle communication between the computer and external devices (keyboards, displays, storage devices)
• Enable data transfer to and from the CPU and memory
• Consist of hardware components (I/O devices, controllers, buses) and software components (device drivers, I/O schedulers)
• Facilitate user interaction with the computer system through input devices and output devices
• Manage the flow of data between the computer and peripheral devices
• Provide an interface for applications to access and manipulate data on storage devices
• Ensure efficient and reliable data transfer while minimizing the impact on system performance
• Abstract the complexities of I/O operations, allowing developers to focus on higher-level functionality

Types of I/O Devices

• Input devices
  • Keyboards and mice for user input
  • Scanners and cameras for capturing images and documents
  • Microphones for audio input
• Output devices
  • Monitors and displays for visual output
  • Printers for producing hard copies of documents
  • Speakers and headphones for audio output
• Storage devices
  • Hard disk drives (HDDs) for high-capacity, non-volatile storage
  • Solid-state drives (SSDs) for faster, more reliable storage
  • Optical drives (CD/DVD) for portable storage and media playback
• Network devices
  • Network interface cards (NICs) for connecting to local area networks (LANs)
  • Modems for establishing internet connectivity
• Specialized devices
  • Graphics processing units (GPUs) for accelerated video rendering and gaming
  • Touchscreens for intuitive user interaction on mobile devices and kiosks

I/O Hardware Basics

• I/O devices connect to the computer system through ports and cables
• Ports provide a physical interface for connecting devices (USB, HDMI, Ethernet)
• Cables transmit data and power between the device and the computer (USB cables, SATA cables)
• I/O controllers manage the communication between the device and the system
  • Implement protocols for data transfer and device control
  • Handle data buffering, error detection, and correction
• Buses facilitate data transfer between components
  • System bus connects the CPU, memory, and I/O controllers
  • Peripheral buses (USB, PCIe) connect I/O devices to the system
• Interrupts signal the CPU when I/O operations require attention
  • Devices raise interrupts to indicate the completion of a task or to request service
  • The CPU suspends its current task and invokes the appropriate interrupt handler

I/O Software Layers

• Application layer
  • Provides high-level APIs and libraries for interacting with I/O devices
  • Enables developers to read from and write to files, communicate over networks, and capture user input
• Operating system layer
  • Manages I/O requests from applications and coordinates access to shared resources
  • Implements file systems, device drivers, and I/O schedulers
  • Provides system calls for applications to perform I/O operations
• Device driver layer
  • Communicates directly with I/O devices and controllers
  • Translates generic I/O requests into device-specific commands
  • Handles device initialization, configuration, and error handling
• Firmware layer
  • Low-level software embedded in I/O devices
  • Implements basic functionality and communication protocols
  • Provides an interface for device drivers to interact with the hardware

Device Drivers and Controllers

• Device drivers are software components that enable communication between the operating system and I/O devices
• Provide a standardized interface for the OS to interact with diverse hardware devices
• Encapsulate device-specific details and present a uniform API to higher-level software
• Responsible for initializing and configuring the device
  • Set up device registers, buffers, and interrupt handlers
  • Allocate necessary resources (memory, DMA channels)
• Handle I/O requests from applications and the operating system
  • Translate generic I/O commands into device-specific instructions
  • Manage data transfer between the device and system memory
• Implement error handling and recovery mechanisms
  • Detect and report device errors (timeouts, data corruption)
  • Attempt to recover from errors or notify the operating system for further action
• Optimize device performance through buffering, caching, and scheduling techniques

I/O Scheduling Algorithms

• I/O scheduling algorithms optimize the order in which I/O requests are serviced to improve system performance
• First-Come, First-Served (FCFS)
  • Processes I/O requests in the order they arrive
  • Simple to implement but can lead to long wait times and poor disk utilization
• Shortest Seek Time First (SSTF)
  • Selects the I/O request with the shortest seek time from the current head position
  • Minimizes disk head movement but may starve requests far from the current position
• SCAN (Elevator)
  • Moves the disk head in one direction, servicing requests along the way, then reverses direction when reaching the end
  • Provides a fair balance between response time and disk utilization
• C-SCAN (Circular SCAN)
  • Similar to SCAN, but moves the disk head in only one direction, then returns to the beginning when reaching the end
  • Ensures a more uniform wait time distribution compared to SCAN
• LOOK and C-LOOK
  • Variants of SCAN and C-SCAN that reverse direction based on the presence of pending requests rather than reaching the end of the disk
  • Avoids unnecessary disk head movement when no requests exist in the current direction

Buffering and Caching in I/O

• Buffering and caching techniques are used to improve I/O performance by reducing the frequency and latency of disk accesses
• Buffering
  • Temporary storage area in memory used to hold data being transferred between devices or between a device and an application
  • Allows I/O operations to be performed asynchronously, enabling the CPU to continue executing while data transfer occurs in the background
  • Smooths out the speed differences between the CPU, memory, and I/O devices
• Caching
  • Stores frequently accessed data in a high-speed memory area (cache) to reduce the need for slower disk accesses
  • Operates on the principle of locality, assuming that recently accessed data is likely to be accessed again in the near future
  • Read caching
    • Retrieves data from the disk and stores a copy in the cache
    • Subsequent reads for the same data can be served directly from the cache, avoiding disk access
  • Write caching
    • Buffers write operations in the cache and periodically flushes the changes to the disk
    • Improves write performance by allowing the application to proceed without waiting for the disk write to complete
• Caching policies
  • Determine which data to cache and when to evict data from the cache
  • Least Recently Used (LRU) evicts the cache entry that has been accessed least recently
  • Least Frequently Used (LFU) evicts the cache entry that has been accessed least frequently

I/O Performance and Optimization

• I/O performance is critical for overall system performance, as I/O operations are often the bottleneck in many applications
• Factors affecting I/O performance
  • Device characteristics (seek time, rotational latency, data transfer rate)
  • I/O request patterns (sequential vs. random access)
  • File system overhead (metadata management, fragmentation)
  • I/O software layers (device drivers, I/O schedulers)
• Techniques for optimizing I/O performance
  • Disk partitioning and file system selection
    • Choose appropriate partition sizes and file systems based on workload characteristics
    • Use journaling file systems (ext4, NTFS) for improved reliability and recovery
  • I/O request merging and reordering
    • Combine adjacent I/O requests into larger, more efficient operations
    • Reorder requests to minimize disk head movement and seek times
  • Asynchronous I/O and non-blocking operations
    • Perform I/O operations asynchronously to allow the application to continue executing while I/O is in progress
    • Use non-blocking I/O APIs to avoid blocking the application when I/O resources are unavailable
  • Caching and buffering
    • Implement effective caching strategies to reduce disk accesses
    • Tune cache sizes and policies based on application requirements and available memory
  • I/O load balancing and parallelization
    • Distribute I/O requests across multiple devices or storage nodes to improve throughput
    • Leverage parallel I/O techniques (striping, RAID) to increase I/O bandwidth
• Monitoring and profiling I/O performance
  • Use system monitoring tools (iostat, iotop) to identify I/O bottlenecks and resource contention
  • Profile application I/O behavior to optimize data access patterns and minimize I/O overhead
  • Analyze I/O traces and logs to identify opportunities for performance tuning and optimization