Operating systems and control systems

A1.3.1 The role of an operating system

An operating system sits between users, applications and hardware. It has two main jobs:

• Abstraction: it hides hardware detail behind simple services, so an app can save a file without knowing how a disk works.
• Resource management: it shares out the CPU, memory, storage and devices efficiently and safely.

Common examples are Windows, macOS and Linux on computers, and Android and iOS on phones.

A1.3.2 Functions of an operating system

Three I/O techniques

Common malware

Next: scheduling →

A1.3.3 Approaches to scheduling

The scheduler decides which process runs next. A pre-emptive scheduler can take the CPU away from a running process. A non-pre-emptive one waits for it to stop.

turnaround = finish − arrival   ·   waiting = turnaround − burst

def fcfs(processes):
    # each process is (name, arrival, burst), sorted by arrival
    clock = 0
    total_wait = total_turn = 0
    for name, arrival, burst in processes:
        clock = max(clock, arrival)     # idle until it arrives
        wait = clock - arrival
        clock += burst
        turn = wait + burst
        total_wait += wait
        total_turn += turn
        print(name, "wait", wait, "turnaround", turn)
    n = len(processes)
    print("average wait", round(total_wait / n, 2))
    print("average turnaround", round(total_turn / n, 2))


fcfs([("P1", 0, 5), ("P2", 2, 3), ("P3", 4, 1), ("P4", 6, 7)])

Next: interrupts and polling →

A1.3.4 Interrupts and polling

Think of a parcel delivery. Polling means you keep walking to the door to check. An interrupt is the doorbell: you carry on working until it rings.

On an interrupt, the CPU saves its state, runs an interrupt service routine (ISR), then restores its state and continues.

Next: multitasking →

Next: review →

HL A1.3.5 Multitasking and resource allocation

Multitasking switches between tasks so quickly that they seem to run at once. The OS must keep this fair, safe and fast.

Mutual exclusion

Only one process may use a shared resource at a time. Otherwise data can be corrupted.

Handling deadlock

Next: control systems →

HL A1.3.6 Control system components

A control system tries to reach a target. The input gives the target, such as a set temperature, and the output is what the system actually does.

Next: real-world uses →

HL A1.3.7 Control systems in use

Each runs as a loop. In an elevator, position sensors keep updating the controller, which slows the motor so that the car stops exactly at the floor. In an autonomous vehicle, transducers convert sensor data into signals and turn commands into action.

int red = 13, yellow = 12, green = 11;

void setup() {
  pinMode(red, OUTPUT);
  pinMode(yellow, OUTPUT);
  pinMode(green, OUTPUT);
}

void show(int led, int ms) {
  digitalWrite(led, HIGH);
  delay(ms);
  digitalWrite(led, LOW);
}

void loop() {
  show(green, 5000);
  show(yellow, 2000);
  show(red, 5000);
}

Next: review →

Quick quiz

Exam-style questions

These are written for this site in the style of Paper 1. Answer on paper first, then open the mark points.

1. Describe how an operating system organizes files. [2]

   Mark points

   • Uses a hierarchical tree of directories containing files [1].
   • Enforces rules such as unique names in a directory, and allows operations such as moving and renaming [1].

2. Outline how an operating system loads an application. [3]

   Mark points

   • Finds the executable on secondary storage [1].
   • Allocates enough memory space for its code and data [1].
   • Copies it into RAM and schedules it to run [1].

3. Describe memory management in an operating system. [3]

   Mark points

   • Allocates and frees RAM for processes as they need it [1].
   • Keeps each process in its own protected space [1].
   • Uses virtual memory on disk when RAM is full [1].

4. Outline why an operating system keeps audit logs. [2]

   Mark points

   • Records activity such as logins and file access [1].
   • Lets administrators detect suspicious activity or diagnose problems [1].

5. Three processes arrive at time 0 in the order A (6 units), B (2), C (1). Calculate the average waiting time under FCFS. [3]

   Mark points

   • A waits 0, B waits 6, C waits 8 [1].
   • 0 + 6 + 8 = 14 [1].
   • 14 ÷ 3 = 4.67 units [1].

6. Compare first come first served and round robin scheduling. [4]

   Mark points

   • FCFS is simple with low overhead [1], but short tasks can wait behind long ones [1].
   • Round robin is responsive and shares the CPU fairly [1], but context switching adds overhead and it depends on the quantum [1].

7. Explain how priority scheduling can leave a process waiting indefinitely, and one solution. [3]

   Mark points

   • High-priority processes keep arriving and take the CPU [1].
   • A low-priority process never gets a turn, which is starvation [1].
   • Ageing raises its priority over time until it runs [1].

8. Discuss whether a hospital heart monitor should use interrupts or polling for a critical alarm sensor. [4]

   Mark points

   • Interrupts respond as soon as the event occurs [1].
   • A critical event is unpredictable and any delay could harm the patient [1].
   • Polling between checks could delay detection [1].
   • Conclusion: interrupts, with the overhead acceptable because alarms are rare [1].