Computer System Organization & Types of Operating Systems

Computer System Organization — The Big Picture

Animated System Diagram — Bus, CPU, Memory, and Controllers

Storage Structure — The Memory Hierarchy

No single storage technology is fast, huge, cheap, and non-volatile all at once. So we build a hierarchy: multiple layers, each trading speed for capacity. The OS is what keeps data flowing between them.

Animated Storage Pyramid — Speed vs Size vs Cost

Practical — Inspect Your Own Memory Hierarchy

# Linux — cache sizes at every level
$ lscpu | grep -i cache
L1d cache:        128 KiB (4 instances)
L1i cache:        128 KiB (4 instances)
L2 cache:         1 MiB   (4 instances)
L3 cache:         8 MiB   (1 instance)

# RAM overview
$ free -h
              total   used   free   buff/cache   available
Mem:           15Gi   6.1Gi  2.4Gi        7.1Gi        8.5Gi
Swap:         4.0Gi   128Mi  3.9Gi

# Storage devices — block layer
$ lsblk
NAME    SIZE TYPE MOUNTPOINT
nvme0n1 512G disk
├─p1    512M part /boot/efi
└─p2  511.5G part /

I/O Structure — How the CPU Talks to Devices

Every I/O device is managed by a small dedicated processor called a device controller. It exposes a set of registers and a local buffer. The CPU never touches the raw device — it talks only to the controller.

Animated Device Controller — CPU, Controller, and Device

3.1 Three Techniques of I/O

Animated Comparison — Polling vs Interrupt vs DMA

3.2 Interrupt Life Cycle

# Linux — watch interrupts happening live per CPU
$ watch -n1 'cat /proc/interrupts | head'
           CPU0    CPU1    CPU2    CPU3
  0:        21       0       0       0   IO-APIC   2-edge      timer
  8:         1       0       0       0   IO-APIC   8-edge      rtc0
 24:    148237   93411   75102   68240   PCI-MSI   nvme0q0    # disk
 25:     22884   31017   27339   29128   PCI-MSI   iwlwifi    # wifi

Types of Operating Systems — Overview

Not all operating systems solve the same problem. Workload, response-time requirement, and hardware shape the design. Galvin lists several major families — we'll cover the full set.

Animated Overview — All OS Types Rotating Around Purpose

Deep Dive — Batch Operating System

Animated Batch Timeline — Serial Execution

Deep Dive — Multiprogramming OS

Batch systems waste the CPU during I/O. Multiprogramming fixes this by keeping multiple jobs in memory at once. When one blocks on I/O, the OS switches to another. The CPU is always doing useful work.

Animated Comparison — Batch vs Multiprogramming

Deep Dive — Multitasking & Time-Sharing OS

Multiprogramming solved CPU utilisation, but it did not care about response time. Users waiting at a terminal want their prompt back in milliseconds. Time-sharing (a form of preemptive multitasking) solves this by giving every process a small time slice.

Animated Time Slice Rotation — Round-Robin Scheduling

# Linux — see scheduling policy and quantum
$ chrt -p 1
pid 1's current scheduling policy: SCHED_OTHER
pid 1's current scheduling priority: 0

# Watch context switches per second
$ vmstat 1
 procs -----------memory---------- --system-- ------cpu-----
  r  b   swpd   free   buff  cache   in    cs us sy id wa
  2  0   131k  2.4G   412M   7.1G  5012 14827  8  3 88  1
                                    ▲      ▲
                             interrupts  context switches

Deep Dive — Real-Time Operating Systems (RTOS)

Animated Deadline Race — RTOS vs General-Purpose OS

Deep Dive — Distributed & Clustered Operating Systems

A distributed OS makes a collection of independent, networked computers appear to the user as a single coherent system. A clustered OS is a close cousin — a group of tightly-coupled machines cooperating for high availability or high performance.

Animated Distributed System — Many Nodes, One Illusion

Clustered vs Distributed vs Network OS — Not the Same Thing

Deep Dive — Embedded & Mobile Operating Systems

Side-By-Side — All OS Types Compared

Practical — Which OS Type Fits Which Problem?

Golden Rules — What to Remember