Operating System Fundamentals

Introduction to Operating Systems

• An operating system (OS) is a program that acts as an intermediary between a user and computer hardware.
• OS goals include executing user programs, solving user problems, making the system efficient to use, and utilizing hardware efficiently.

Computer System Components

• Hardware provides the fundamental computing resources (CPU, memory, I/O).
• The OS controls and coordinates hardware use among application programs for users.
• System and application programs define how system resources are used to solve user computing problems like compilers, video games or database systems.
• Users can be people, machines, or other computers.

Operating System Definitions

• A resource allocator manages and allocates resources.
• A control program controls the execution of user programs and I/O device operations.
• The kernel is the core program running at all times; all else is system and application programs.

Multiprogramming Batched Systems

• Multiprogramming keeps several jobs in main memory at once, multiplexing the CPU among them, which requires memory management and protection.
• The OS picks and executes jobs from memory, switching to another job when the current one needs an I/O operation. This keeps the CPU and OS consistently active.
• The number of jobs in memory is less than the number of all jobs on disk.
• If memory is limited, the system chooses which jobs to bring in, in other words job scheduling.
• If several jobs are ready to run, the system selects among them, in other words CPU scheduling.
• Running multiple programs simultaneously requires memory management.
• In older systems, the CPU sits idle when waiting for I/O. Conversely in multiprogramming, the CPU is continuously active.

OS Features Needed for Multiprogramming

• OS Features needed for Multiprogramming include I/O routines, memory management, job and CPU scheduling, device allocation, and protection between the OS and programs.
• Two main disadvantages of multiprogrammed batched systems include the user being unable to interact with their jobs while executing, and modification prohibition to study behavior.

Time-Sharing Systems - Interactive Computing

• A time-sharing system uses CPU scheduling and multiprogramming to give each user a small portion of a time-shared computer.
• The CPU is multiplexed among jobs in memory and on disk, and is allocated to a job only if the job is in memory.
• The CPU switches jobs when one needs I/O, ensuring it remains consistently active.
• Jobs are swapped between memory and disk, with the disk providing backup for main memory.
• Online communication is provided between the user and the system.

Time-Sharing Systems

• Disk management is a necessity for online file systems, because they store the files on collections of disks.
• Concurrent execution uses CPU scheduling schemes.
• Job synchronization and communication are required for orderly execution.
• A mechanism to avoid deadlock is needed, to prevent jobs from waiting for each other indefinitely.

Personal-Computer Systems

• Personal computers are dedicated to a single user.
• They use I/O devices like keyboards, mice, display screens, and small printers.
• PC operating systems were designed without multi-user or multi-tasking capabilities.
• PC operating systems prioritize user convenience and responsiveness over maximizing CPU and I/O utilization.
• Examples of PC operating systems are Microsoft Windows and Apple Macintosh.

Parallel Systems

• Parallel systems are multiprocessor systems with more than one CPU in close communication.
• Tightly coupled systems involve processors sharing memory and a clock with the communication often taking place across the shared memory.

Parallel system advantages

• Increased throughput improves the number of processes completing per time unit.
• Economics improve for large jobs, eliminating the need to make copies and distribute them among processors.
• Speed increases for large jobs, because work is divided across processors.
• Reliability increases via fault tolerance; when a processor fails, the remaining ones pick up the slack.

Parallel System Models

• Symmetric multiprocessing (SMP) has each processor running an identical copy of the OS.
• Most modern operating systems support SMP, like UNIX for the Multimax computer.
• Asymmetric multiprocessing has each processor assigned a specific task, with a master processor scheduling and allocating work to slave processors.
• The master performs I/O and computations, and only it can execute the OS.
• Slaves can only execute user programs; the system cannot perform I/O if the master fails.
• If a slave fails, it will cause some computations to be lost but the system will still function.
• Asymmetric multiprocessing is more common in extremely large systems.

Distributed Systems

• Distribute computation among multiple physical processors.
• Loosely coupled systems connect two or more independent computer systems via a communication link.
• Each processor in Loosely coupled systems has its own OS and local memory.
• Processors communicate through lines (message passing) like high-speed buses or telephone lines.
• Shared files and printers provide resource sharing.
• Computation speed increases, because a job can be partitioned so each processor concurrently does a portion in load sharing.
• Reliability increases because if one processor fails the rest still function.
• Communications increases with things such as electronic mail and ftp.
• LANs (Local Area Network), WANs (Wide Area Network), and MANs (Metropolitan-Area Network) are types of distributed systems.
• Compute-servers provide an interface to clients for requesting services like database access.
• File-servers provide an interface for clients to store and retrieve files.

Client-Server and Peer-to-Peer Systems

• Client-server systems involves a network connecting multiple clients to a central server.
• Peer-to-peer (P2P) systems do not distinguish between clients and servers with all nodes considered peers that can act as client, server, or both.
• A node in P2P joins the P2P network, registers its service with a lookup service, and responds to service requests via a discovery protocol.
• Napster and Gnutella are examples of P2P systems.

Clustered Systems

• Clustered systems are like multiprocessor systems but involve multiple systems working together, often sharing storage via a storage-area network (SAN).
• They provide high-availability by surviving failures through:
• Asymmetric clustering (one machine in hot-standby mode).
• Symmetric clustering (multiple nodes running applications, monitoring each other).
• Some clusters support high-performance computing (HPC), with applications written to use parallelization.

Real-Time Systems

• Real-Time Systems provide immediate responses, such as sensors bringing the data directly into the computer.
• Often used as control devices in specialized applications, like scientific experiments, medical imaging or industrial control systems, and some display systems.
• Well-defined fixed-time constraints are a must.
• Hard real-time systems have limited or absent secondary storage, with data stored in short-term or read-only memory (ROM).
• Hard real-time systems have conflicts with time-sharing systems and are not generally supported by general-purpose OSs.
• Soft real-time systems have limited utility in industrial control or robotics, but are useful in multimedia or virtual reality applications needing advanced OS features.

Computer-System Architecture

• I/O devices and the CPU can execute concurrently.
• Each device controller manages a specific device type and has a local buffer.
• The CPU moves data between main memory and local buffers.
• I/O is done from device to the local buffer of controller.
• The memory controller ensures orderly shared memory access by synchronizing access to the shared memory.
• A device controller informs the CPU about operation completion by causing an interrupt.

Storage Structure

• Main memory is the only directly accessible storage media for the CPU; programs must be in it to execute.
• Secondary storage extends main memory, providing large, nonvolatile storage capacity.
• Magnetic disks are metal or glass platters covered with magnetic recording material.
• Disks are logically divided into tracks subdivided into sectors.
• Disk controllers determine the logical interaction between the device and the computer.
• The two surfaces of a platter are covered with magnetic material.
• Platter diameters range from 1.8 to 5.25 inches.
• There may be thousands of concentric cylinders in a disk drive, and each track may contain hundreds of sectors.
• The transfer rate is the rate at which data is transferred between the drive and the computer.
• Positioning time (or random access) consists of seek time and rotational latency.
• Seek time is the time to move the disk arm to the desired cylinder.
• Rotational latency is the time for the desired sector to rotate to the disk head.
• Floppy disks consist of one platter with the head directly sitting on the surface. It’s inexpensive, has 1.4MB storage, and is slow compared with the hard disk.
• The load instruction moves a word from main memory to an internal register to be executed.
• The store instruction moves the content of a register to main memory.
• Main memory is too limited to permanently store all the data, so secondary storage is used.
• RAM is volatile, losing contents once power is off or lost.

Storage Hierarchy

• Storage systems are organized in a hierarchy based on speed, cost, size, and volatility.
• Caching copies information into faster storage, so main memory can be considered a last cache for secondary storage.
• Registers, cache, and main memory are volatile.
• All storage after main memory is non-volatile.
• Caching involves checking cache memory first and then copying data from main memory assuming it will be needed again.
• Data must be moved from secondary storage into main memory before use.
• Data transfer from cache to CPU and registers is usually in hardware without OS involvement.
• Data transfer from disk to memory is controlled by the OS.
• Cache Coherency and Consistency in multiprocessor systems means that if shared data is in 2 or more caches, no caches have different values simultaneously.

Hardware Protection

• Dual-Mode Operation provides hardware support to differentiate between at least two operation modes: user mode and monitor mode.

Dual-Mode Operation

• Sharing system resources demands the operating system makes sure an incorrect program doesn't cause other programs to execute incorrectly.
• User mode is for executing user programs on behalf of a user.
• Monitor mode, also known as supervisor/system mode, operates on operating system's behalf.
• At boot time, hardware starts in monitor mode. The OS loads, then starts user processes in user mode.
• When an interrupt occurs, hardware switches from user to monitor mode.
• When the OS gains control, it is in monitor mode
• If dual mode is absent, systems are subject to the dual mode problem
• Examples: MS-DOS for 8088 architecture lacks a dual mode, and MS-Windows NT and IBM OS/2 utilize the dual mode for greater protection.
• A mode bit is added to the computer hardware to indicate current mode: Monitor (0) or User (1).
• When an interrupt or fault occurs, hardware switches to monitor mode.
• Privileged instructions can only be issued in monitor mode.

I/O Protection

• All I/O instructions are privileged instructions.
• A user program can not assume control of the computer while it is in monitor mode.

User of a System Call to Perform I/O

• The system can execute I/O instructions only by the O.S.
• For I/O, a user executes a system call requesting that the O.S. perform I/O.

Memory Protection

• The goal is to protect the OS from user programs, and isolate user programs.
• In cases of memory protection, two registers are added to establish the legal address range.
• A base register holds the smallest legal physical memory address.
• The Limit register contains the size of the legal range.
• Memory outside the defined range is protected.
• This hardware protection is accomplished in the CPU that only compares addresses generated that are in User Mode with the registers.
• The base/limit registers can only be loaded by the O.S.