第6章インフラストラクチャーの抽象化層：オペレーティングシステム

Now that you’ve been introduced to the basics of operating systems, it may come as no surprise to you that operating systems come in all shapes and sizes. There are many different types of OSs (from batch operating systems to network operating systems), and each type handles many different functions (from file management to device management). This OS tutorial (https://openstax.org/r/76OStutorial) explains the different types. As you read along, consider how much you interact with an OS over the course of a single day—or how many different OSs you interact with.

The kernel is the program that is running at all times on the computer and provides basic services for all parts of the OS. Typically, the kernel is not the only program that is running; there is usually either a system program/operating system or an application program running as well. Any program running on top of the OS is considered to be a process from the OS standpoint. It includes the program code, its current activity represented by the program counter, and a set of resources like open files and allocated memory. It also consists of address space, one or more threads of control executing in that address space, and an additional system state associated with it. The operating system uses processes to manage the execution of programs, ensuring they have the necessary resources while keeping them separate from each other to prevent interference.

Software applications have to be compiled and linked with system libraries before they can run as executable programs on a machine. Each running program runs in its own process, and the OS can run, switch, and isolate processes from each other even though they are actually running on the same hardware. For a given application, the “machine” is the process abstraction provided by the OS. Processes provide user-friendly interfaces, rather than raw hardware, and an execution environment with restricted rights controlled by the OS.

An OS virtualizes the machine by providing easy-to-use abstractions of physical resources while masking limitations. In this context, a thread is a path of execution within a process, and a process may contain multiple threads. Thus, multithreading involves executing multiple threads (i.e., execution units that are part of a process and share the same resources) concurrently, which improves overall responsiveness and efficiency.

In the fast-moving world of IT, it is crucial to think globally. This global perspective can be particularly useful when considering how IT and OSs affect health care. Namely, it can help us understand the role of OSs in patient care, medical research, and the efforts being made to make health care more widely accessible. The challenges associated with data sharing, privacy, and health-care disparities impact regions around the world differently.

As the delivery of health care becomes more integrated with technology, the choices made in designing how health-care IT systems use OSs can significantly impact people’s lives across the globe. These choices range from deciding how to take ethical considerations into account when using AI for diagnosis to creating IT solutions that respect cultural differences.

Check out these initiatives from the World Health Organization (https://openstax.org/r/76WHO) and Centers for Disease Control and Prevention (https://openstax.org/r/76CDC) for examples in the health-care industry.

Imagine you’re considering entering the health-care IT field. Can you think of some ways that thinking globally and understanding the capabilities provided by OSs could improve your ability to solve problems and create solutions that work well for health-care systems in various countries?

The ISA defines how the CPU is controlled by the software by abstracting the hardware details from the applications. The OS provides an abstract machine interface to the application programs and leverages the physical machine interface to do so. The OS communicates with input/output (I/O) hardware using device drivers, I/O ports, interrupts, direct memory access (DMA), and effective I/O scheduling. It provides abstractions to manipulate files (i.e., streams) and send messages to the network (i.e., sockets). Programming languages provide application programming interfaces (APIs) that leverage these abstractions (e.g., file I/O, socket libraries, and related APIs) so that the application program can access the underlying resources that are managed by the OS. One of the main responsibilities of the OS is to isolate hardware from programs by providing common services and background management functionality, for example, storage manager, network manager, and power manager.

The OS is the only system that should be able to directly access I/O devices (i.e., disks, network cards) and manipulate memory. Moreover, the CPU hardware provides a privileged instruction that can only be executed by the OS. The OS can use these instructions to establish an execution environment that limits access (to, for example, memory). The application cannot remove the restrictions because it must execute privileged instructions to do so.

Certain operations are prohibited when running in user mode, such as changing the page table pointer (i.e., the pointers to memory pages that are cached for faster access), disabling interrupts (i.e., the interrupts that the OS received from I/O devices), interacting directly with hardware, and writing to kernel memory. Carefully controlled transitions between user mode and kernel mode include system calls, interrupts, and exceptions. The system call appears when the program requests a service from the kernel. The system interrupt manages the communication between the computer hardware and the system. The system throws an exception, which is an error that occurs at runtime.